JOURNAL OF INFORMATION SYSTEMS & OPERATIONS MANAGEMENT

VOL. 19 NO. 1
MAY 2025

ROMANIAN-AMERICAN UNIVERSITY PUBLISHING HOUSE

BUCHAREST
 EDITOR-IN-CHIEF NOTE

JISOM is a journal which enables researchers, academia, and professionals to make

their voices heard, to share their findings with their peers and the public at large.
Being an open-access journal, JISOM aims to increase the all-around level of
knowledge in the fields of computer science and economics, to further the society’s
capabilities to understand new concepts, to see how things are done by cutting-edge
technologies implementations, to understand what is in store for us not only at the
present time but also in the near, and so near, future.
The published articles focus mainly on IT&C but we also provide a favorable
exposure medium for correlated topics, such as economics, management, applied
sciences, mathematics, statistics, etc. JISOM encourages cross-disciplinary research
of national and international researchers and welcomes the contributions which give
new and fresh perspectives to the above-mentioned fields of study.
National and international researchers, professionals, recognized experts, professors
who want to share their research results and new ideas, and Ph.D. students who want
to improve their knowledge or present their emerging doctoral research are all
welcome to join our knowledge-eager community.

We have a history that started back in 2007 and we are permanently striving to bring
our community to a higher level of knowledge in the fields we cover. With the
constant support of our authors, reviewers, readers, and editorial staff I am sure we
are and will live up to the mission we have taken on.

Many thanks to our JISOM community and good luck with your research!

Respectfully,
Alexandru TĂBUȘCĂ, PhD
JISOM Editor-in-Chief
 JOURNAL OF INFORMATION SYSTEMS & OPERATIONS MANAGEMENT

EDITOR-IN-CHIEF
Alexandru TĂBUȘCĂ, PhD Associate Professor

MANAGING EDITOR
Daniela Alexandra CRIȘAN, PhD Associate Professor

EDITORIAL BOARD
Academician Gheorghe Păun Romanian Academy, Romania
Professor Alexandru Pîrjan Romanian-American University, Romania
Professor Allan Berg University of Dallas, USA
Professor Cornelia Botezatu Romanian-American University, Romania
Professor Costin Boiangiu Politehnica University, Bucharest, Romania
Professor Carlos Rompante da Cunha Bragança Polytechnic University, Portugal
Professor Cristina Coculescu Romanian-American University, Romania
Professor Eduard Rădăceanu Romanian Technical Academy, Romania
Professor George Căruțașu Romanian-American University, Romania
Professor Ion Ivan Academy of Economic Studies, Romania
Professor Ion Smeureanu Academy of Economic Studies, Romania
Professor Kent Zimmerman James Madison University, USA
Professor Lucia Rusu Babes-Bolyai University, Romania
Professor Pauline Cushman James Madison University, USA
Professor Ramon Mata-Toledo James Madison University, USA
Professor Sergiu Iliescu Politehnica University, Bucharest, Romania
Professor Traian Muntean Universite Aix–Marseille II, France
Professor Victor Patriciu National Technical Defence University, Romania
Professor Viorel Marinescu Technical Civil Engineering, Romania
Associate Professor Alexandru Tăbușcă Romanian-American University, Romania
Associate Professor Irina Făgărășanu Politehnica University, Bucharest, Romania
Associate Professor Mihaela Păun Louisiana Tech University, USA
Associate Professor Sanda Micula Babes-Bolyai University, Romania
Associate Professor Susan Kruc James Madison University, USA

Senior Staff Text Processing:

Lecturer Gabriel Eugen Garais Romanian-American University
Lecturer Justina Lavinia Stănică Romanian-American University
 JOURNAL OF INFORMATION SYSTEMS & OPERATIONS MANAGEMENT
Journal details

No. Item Value

JOURNAL OF INFORMATION
1 Complete title / IDB title SYSTEMS & OPERATIONS
MANAGEMENT

2 ISSN (print and/or electronic) 1843-4711

3 Frequency semestrial

Journal website (direct link to journal

4 http://jisom.rau.ro
section)

EBSCO

GALE Cengage Learning

5 IDB indexation Index Copernicus

ProQuest

RePEC/IDEAS

Contact

First name and last name Alexandru TĂBUŞCĂ,

PhD Associate Professor

Phone +4-0372-120.140

E-mail alex.tabusca@rau.ro

ISSN: 1843-4711
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

TABLE OF CONTENTS

THE ACCOUNTING SERVICES QUALITY MANAGEMENT IN THE CONTEXT OF

ACCELERATED DIGITALIZATION .......................................................................................... 1
LUCIAN CONSTANTIN GABRIEL BUDACIA
ELISABETA ANDREEA BUDACIA
MARIAN FLORIN BUSUIOC
BUILDING A SUSTAINABLE FUTURE BY DEVELOPING THE CONCEPT OF DOUBLE
DIAMOND IN TRIPLE DIAMOND, A NEW CONCEPT IN CREATING A NUDGE
TOWARDS PRO-ELECTRIFICATION BEHAVIOR IN URBAN FREIGHT TRANSPORT
......................................................................................................................................................... 11
IRINA CALOTĂ
ALEXANDRA PERJU-MITRAN
AUGUSTIN SEMENESCU
RETHINKING HR MANAGEMENT FOR THE DIGITAL WORKPLACE ......................... 27
NICULINA CHIVU
GEORGE CARUTASU
MEASURING THE PERFORMANCE OF A SMART HOME AUTOMATION SOFTWARE
USING DESIGN PATTERNS....................................................................................................... 45
TUDOR-CĂLIN CIOT
ŞTEFAN BUTACU
COSTIN-ANTON BOIANGIU
CĂTĂLIN TUDOSE
ONLINE VERSUS IN-PERSON PRODUCTIVITY IN DIFFERENT TEAMWORK
CONTEXTS ................................................................................................................................... 60
ALEXIA CIUCLEA
ANDREEA-BIANCA ŞTEFAN
ŞTEFAN STAN
COSTIN ANTON BOIANGIU
CĂTĂLIN TUDOSE
FACILITATING ITALIAN LANGUAGE ACQUISITION AMONG ROMANIAN
SPEAKERS IN AN ONLINE SETTING: THE ADVANTAGES OF INTEGRATING DIGITAL
RESOURCES ................................................................................................................................. 74
MARIANA COANCĂ
EXPLORING THE LINK BETWEEN INNOVATION AND ICT READINESS. A
COMPARATIVE ANALYSIS OF ESTONIA AND ROMANIA ............................................... 89
DANIELA ALEXANDRA CRIȘAN
JUSTINA LAVINIA STĂNICĂ
USABILITY TESTING APPLIED TO GRAPHICAL AND VR APPLICATIONS IN AN
ACADEMIC ENVIRONMENT ................................................................................................. 103
CRINA DUTA
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

NICOLETA LUMINITA CĂRUȚAȘU

GEORGE CĂRUȚAȘU
IONUȚ-CRISTIAN PREDERIC
QUANTITATIVE AND QUALITATIVE DIGITAL ANALYSIS OF EMOTIONAL AND
MOTOR INTELLIGENCE BY GENDER DIFFERENCES................................................... 116
IOANA GABRIELA GRIGORESCU
EUGEN GABRIEL GARAIS
CHALLENGES, OPPORTUNITIES AND IMPLICATIONS REGARDING THE
INTEGRATION OF ARTIFICIAL INTELLIGENCE IN AUDIT PROCESSES ................. 138
MIHAELA PANAIT (ION)
MARILENA-ROXANA ZUCA
AURA OANA MUSTĂȚEA
VICTOR MUNTEANU
CHANGES IN BEHAVIOR OF UNIVERSITY STUDENTS AFTER COVID-19 SELF-
RESTRAINT PERIOD ............................................................................................................... 162
SHOGO KOYAMA
TAKUNE SAKAUE
NOBUTAKA SUZUKI
ARTIFICIAL INTELLIGENCE SOLUTIONS FOR ENERGY CONSUMPTION
OPTIMIZATION IN IOT DEVICES ........................................................................................ 178
RĂZVAN MOCANU
FLORENTINA NIDELCU
GEORGE CĂRUȚAȘU
THE DIGITAL TRAP: TEENS AND ONLINE CHALLENGES ........................................... 196
RADU MOINESCU
CIPRIAN RĂCUCIU
CARMEN-SILVIA OPRINA
DRAWING ON AUTOCAD THE HORN OF THE ICONIC ALTEC LANSING
LOUDSPEAKER VOICE OF THE THEATER A7 ................................................................. 211
JOSE MUJICA
GENERATIVE PRETRAINED TRANSFORMERS FOR INVESTOR-CENTRIC
PORTFOLIO CONSTRUCTION .............................................................................................. 226
DIMITRIOS PAPAKYRIAKOPOULOS
MANOLIS KRITIKOS
SMART HOSPITALITY: THE ROLE OF AI IN ENHANCING SUSTAINABLE GUEST
EXPERIENCES ........................................................................................................................... 241
IOANA CRISTIANA PATRICHI
ARTIFICIAL SOCIAL INTELLIGENCE AND THE TRANSFORMATION OF HUMAN
INTERACTION BY ARTIFICIAL INTELLIGENCE AGENTS ........................................... 259
ALEXANDRU PÎRJAN
DANA-MIHAELA PETROŞANU
GENERATING JAVA CODE WITH AI TOOLS. USAGE AND IMPLICATIONS .............. 306
ALEXANDRU TĂBUȘCĂ
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ANDREI LUCHICI
MIHAI BOTEZATU
SILVIA TĂBUȘCĂ
FINANCIAL ACCOUNTING INFORMATION IN THE CONTEXT OF SCORE-BASED
SUSTAINABILITY REPORTING ESG (ENVIRONMENTAL, SOCIAL, GOVERNANCE)
....................................................................................................................................................... 329
MARILENA ZUCA
ALICE EMILIA ȚÎNȚA
ANDA MIHAELA LĂCEANU
GEORGIANA POPA
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

THE ACCOUNTING SERVICES QUALITY MANAGEMENT IN THE CONTEXT

OF ACCELERATED DIGITALIZATION

Lucian Constantin Gabriel BUDACIA1

Elisabeta Andreea BUDACIA2
Marian Florin BUSUIOC3

Abstract
In the field of accounting, the concept of “quality” refers to the extent to which the services
provided comply with professional standards, legal regulations, customer expectations and
organizational requirements. We have identified a series of particular dimensions of
accounting service quality that we have developed in this article. Quality management in
this context is not limited to compliance with technical norms, but also includes factors
such as customer satisfaction, process efficiency and adaptability to external changes.
Quality in accounting services plays an essential role, having a direct impact on the
financial credibility of an organization and its relations with tax authorities and
stakeholders. Accounting is a practical activity carried out by professionals, which has the
role of providing useful information for decision-making at the level of economic entities.
Every user of accounting information wants to have information that reflects reality, that
is, true information.
Keywords: quality of accounting services, dimensions of quality, quality management,
quality principles
JEL Classification: L15, M29, M41

1. Introduction
In the field of accounting, the concept of "quality" refers to the extent to which the services
provided comply with professional standards, legal regulations, customer expectations and
organizational requirements. Quality management in this context is not limited to
compliance with technical norms, but also includes factors such as customer satisfaction,
process efficiency and adaptability to external changes. Quality in accounting services plays

1 PhD, Junior Lecturer, Romanian – American University, lucian.constantin.budacia@rau.ro

2 PhD, Senior Lectuer, Romanian – American University, andreea.budacia@rau.ro, corresponding author
3 PhD, Senior Lectuer, Romanian – American University, marian.florin.busuioc@rau.ro

Pag. 1 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

a crucial role, having a direct impact on the financial credibility of an organization and its
relations with tax authorities and stakeholders.
Considering the concept of specialist J. M. Juran [1], according to which quality
management has three pillars: planning, control and improvement, we can notice a series
of their particularities regarding accounting services.
■ Quality planning is the activity of developing the processes necessary to satisfy customer
needs. The stages of quality planning consist of establishing the following elements: setting
quality objectives; identifying customer characteristics; determining customer needs;
designing the characteristics of works/processes that meet customer needs; designing/using
technological tools that produce the requested characteristics; establishing control methods
for the performance process.
■ Quality control establishes the infrastructure necessary for assessing real quality and the
possibility of comparing it with the quality objectives of the process by acting on the
differences between real and achieved quality.
■ Quality improvement is the process of improving performance by identifying projects
and ensuring the necessary infrastructure for diagnosing causes and finding remedies and
establishing the resources necessary for quality improvement.
Outlining the characteristics of services is an important aspect, which allows their
individualization from other elements that are the object of exchanges within the market.
The opinions of specialists differ regarding the characteristics of services, considered an
important issue both from a conceptual and practical point of view. The following general
aspects are mainly taken into account [2]: intangibility, inseparability, variability,
perishability.
Intangibility makes it very difficult for the provider to convince the client to use the
respective services. In practice, the beneficiary may have some reservations about the
capacity of the respective services to satisfy their organizational needs.
Inseparability highlights the fact that services cannot be separated from their provider
either spatially or temporally. In practice, we are talking about the simultaneity of
production/performance and consumption.
Variability highlights the fact that the quality of services depends on the provider, when,
how and where they provide them, which makes a service differ from one performance to
another. A service can almost never be repeated in an identical way. Often, customers
negatively assess the differences between the services they receive, because discrepancies
appear between the quality promised, delivered, and received.
Perishability basically shows that services are not storable.

2. The dimensions of the quality of accounting services

Pag. 2 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Starting from these general characteristics, we can identify a number of specific dimensions
of the quality of accounting services.
Correctness
It is essential that accounting services are correct, complying with the accounting and tax
regulations in force. This aspect includes carefully checking financial documents and
ensuring that all calculations and reports are carried out correctly.
Accuracy
Accuracy is fundamental to maintaining the trust of clients and tax authorities, as
accounting errors can lead to financial penalties.
Clarity
Accounting services must be clear and transparent, allowing clients to easily understand
financial reports, balance sheets and other documents.
Confidentiality
Accounting services must ensure that information is protected from unauthorized access
and is managed in accordance with data protection regulations.
Communication
Quality in accounting services also involves effective communication with clients. This
involves speed and clarity, accurate information, providing regular updates and being
available to address client issues.
Professionalism
Professionalism in accounting services involves adhering to a strict code of ethics, as well
as maintaining appropriate behavior towards clients, colleagues and tax authorities. This
also includes behaviors related to integrity, transparency and accountability. Ethical and
professional behavior is crucial for maintaining reputation and trust.
Process efficiency
Accounting services must be efficient in terms of resource use (time, people, technology)
and minimize costs, while maintaining a high level of quality. The efficiency of accounting
processes contributes to reducing errors, saving time and improving client satisfaction.
Adaptability
Quality accounting services are able to adapt to frequent changes that may be legislative,
economic or technological. Adaptability allows an accounting firm to remain competitive
in the market, and innovation can improve services and reduce costs.

Pag. 3 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3. The fundamental principles of the Quality Management

According to the ISO 9001:2015 standard on quality management systems, there are 7
fundamental principles that form the basis for the implementation and continuous
improvement of a quality management system. These principles are [5]:
A. Customer focus
B. Leadership
C. Engagement of people
D. Process approach
E. Improvement
F. Evidence-based decision-making
G. Relationship management

3.1 Customer Focus

Organizations must understand and meet customer requirements and expectations.
Customer satisfaction is essential to the success of the organization, and continuous
improvement of customer satisfaction must be a priority. Thus, quality management must
be focused on meeting customer needs and requirements. In the field of accounting, this
means that services must be tailored to the specifics of each client and meet their
expectations, while ensuring transparency, accuracy and clarity. Clients are the main source
of income for accounting firms. Therefore, client satisfaction is a key indicator of service
quality. A customer-oriented approach will contribute to increasing customer loyalty and a
positive reputation for the firm. Conducting periodic assessments of customer satisfaction,
integrating their feedback into the service improvement process and personalizing
accounting services according to the client's needs.

3.2 Leadership
Leaders must set the purpose and direction of the organization. They must create a favorable
internal environment for employees to be involved in achieving the organization's goals.
Leadership refers to the ability to guide, inspire, and influence a group of people to achieve
common goals. An effective leader not only makes important decisions, but also motivates,
supports, and develops teams to perform at the highest level. Leadership also involves
accountability, integrity, and vision. In the field of accounting, leadership manifests itself
in several ways:
■ Managing Accounting Teams:
Pag. 4 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Accounting leaders must coordinate teams, assign tasks, and ensure that all members are
effective and understand the importance of their work. They are responsible for developing
and implementing accounting strategies.
■ Strategic Decisions:
An accounting leader must make decisions that ensure not only the organization’s financial
compliance but also its long-term financial health. This may include tax planning, assessing
financial risks, or adopting accounting best practices.
■ Promoting an ethical and compliant environment:
A leader in the field must be an example of integrity, respecting professional standards and
tax and financial regulations. He must ensure that the team complies with accounting
regulations, avoiding any form of fraud or unethical practices.
■ Innovation and adaptability:
In a constantly changing field such as accounting, a leader must be able to identify new
technologies and methodologies that can improve the efficiency of accounting teams and
respond quickly to legislative or economic changes.
■ Training and professional development:
Leadership in the accounting field also includes supporting the professional development
of teams. Leaders must encourage continuous training, provide constructive feedback, and
help team members improve their skills and competencies.

3.3 Engagement of people

The involvement of all employees is essential for the success of the organization. People
must be involved and motivated to actively contribute to achieving the organization's
objectives, and their skills must be constantly developed. Quality management should not
be the exclusive responsibility of a department or a group of employees, but must involve
the entire organization. Each employee, regardless of their hierarchical level, must
contribute to maintaining quality standards within the organization.
In an accounting firm, all employees, from accountants to managers, must understand and
apply the principles of quality management. The involvement of all ensures team cohesion
and alignment to the same objectives. Creating a work environment that encourages the
active participation of all employees in quality improvement activities, their continuous
training and encouraging open and constructive feedback.

3.4 Process Approach

Pag. 5 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Organizing activities and resources in a systematic and efficient manner is important for
achieving consistent and predictable results. The process approach ensures better
management of risks and opportunities.
In accounting, this involves analyzing and optimizing workflows and procedures to ensure
that all activities are carried out efficiently and correctly. A process-based approach allows
for more efficient management of accounting activities, helping to reduce errors and
processing time, as well as improving transparency and consistency. Identifying and
optimizing key processes in accounting (for example, the process of recording financial
transactions, preparing financial reports, managing payments and collecting taxes and
automating them), where possible and aligning them with internal quality standards.

3.5 Improvement
Improvement is a permanent goal of the organization. A constant focus on processes and
performance helps to identify and implement continuous improvement measures, which
leads to better efficiency and effectiveness. In the field of accounting, this involves
constantly evaluating the efficiency and effectiveness of accounting processes and
identifying ways to optimize them.
The accounting market is dynamic and subject to frequent legislative and economic
changes, and improvement ensures the company's adaptability to these changes and
maintaining a quality service in the face of external challenges.
This principle is applicable through the implementation of improvement cycles based on
the Plan-Do-Check-Act (PDCA) model, periodic evaluation of internal processes and the
adoption of innovative solutions, such as the automation and digitalization of accounting
activities.

3.6 Evidence-Based Decisions Making

Quality management is based on data-based decision-making and systematic performance
analysis. Decisions should be made based on objective analysis and evidence. The proper
organization and use of data helps to make informed decisions and optimize processes.
In accounting, this principle is applied by using financial and statistical data to evaluate the
efficiency and accuracy of accounting processes. Informed decisions lead to more efficient
and accurate solutions, and in the accounting field, this can reduce errors and risks, thus
improving the quality of services.
This principle is applicable by using advanced accounting software, implementing a
performance management system (e.g., KPIs – key performance indicators), and collecting
relevant data to evaluate internal processes and performance.
Pag. 6 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3.7 Relationship Management

In the context of quality management, this principle refers to the creation and maintenance
of strong and trusting relationships between all stakeholders – employees, customers,
suppliers and tax authorities. In the accounting field, close collaboration with tax
authorities, customers and other external parties is essential to ensure a quality service.
Collaboration with them can lead to increased performance and mutual satisfaction.
Partnership relationships allow for better coordination and communication between the
parties involved and ensure that all parties work towards a common goal: improving service
quality and customer satisfaction.
This principle is applicable by building trusting relationships with customers and tax
authorities through transparency and compliance with commitments, collaborating with
technology providers to improve IT infrastructure and ensuring clear and effective
communication with all stakeholders.

4. Challenges in implementing quality management in accounting

Implementing a quality management system in accounting services can be a complex
process, requiring significant obstacles to overcome [4]. These challenges are related both
to the specific nature of the accounting field and to external factors, such as constantly
changing regulations and customer requirements. The main challenges encountered in
implementing quality management in this sector are presented below.
● Complexity of regulations and frequent legislative changes
The accounting field is regulated by norms and standards that change frequently, which can
make it difficult to maintain a consistent quality practice. It should also be borne in mind
that the tax code also undergoes changes, additions and additions annually. Therefore,
accounting firms must always be aware of these changes in order to comply with current
regulations. This can lead to difficulties in ensuring consistent quality, especially in the
context where the team needs constant training in the context of frequent changes.
● Lack of process standardization
In many accounting firms, processes are not always standardized, which is why the services
provided can vary significantly from one client to another. Also, many accounting activities
rely on the individual experience of accountants, which can lead to inconsistencies in the
application of procedures. Without process standardization, it is difficult to maintain a
consistent level of quality. Accounting firms can experience difficulties in providing
consistent and predictable services, which affects both efficiency and client satisfaction.
● Resistance to change and adaptation to new technologies
Pag. 7 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Implementing quality management often involves changes in internal processes, adopting

new technologies and working methods. Many accounting professionals can be reluctant to
change, which can delay the implementation of a quality management system and make it
difficult to integrate new technologies, such as advanced accounting software, automated
audit systems or blockchain technologies.
● Limited resources of small firms
Many accounting firms, especially medium and small ones, have limited financial and
human resources to invest in implementing a complete quality management system.
The lack of resources can make it difficult to implement effective quality processes, from
training staff to purchasing modern technologies to support service quality.
● Complexity of human resource management
Continuous training of staff and maintaining a high level of competence can be a challenge.
Staff turnover can affect the continuity of quality processes. Also, differences in experience
and training among team members can lead to variations in the quality of services provided.
● Measuring and evaluating service quality
Measuring quality in accounting is more complex than in other service areas, as the success
of an accounting service cannot always be assessed directly and immediately. Establishing
performance indicators that reflect the quality of accounting services and can be measured
objectively is a challenge. For example, client satisfaction or performance before tax
authorities are factors that can vary greatly depending on the context.
● Confidentiality and information security
Accounting services involve the management of a large number of sensitive client data, and
their protection is an absolute priority. Failure to comply with confidentiality regulations
can seriously damage reputation and lead to legal sanctions. In implementing a quality
management system, accounting firms must ensure that financial information is properly
managed and protected, which may involve significant investments in IT security systems
and staff training.
● Adapting to diverse client requirements
Each client has different needs and varying expectations regarding accounting services.
This can sometimes lead to difficulties in delivering a uniform and consistent quality
service. Managing these diverse requirements can lead to difficulties in establishing unified
internal procedures, which can affect the efficiency and consistency of the services
provided.

Pag. 8 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

5. Conclusions
Quality in accounting services is not limited to the technical accuracy of financial
documents, but involves a holistic approach that includes transparency, client satisfaction,
professionalism and process effectiveness. In a constantly changing economic environment
with complex regulations, accounting firms must implement quality management methods
that respond to both external requirements and client needs to ensure long-term success.
The principles of quality management in accounting are fundamental to creating an efficient
work environment and providing high-quality accounting services. By implementing client-
oriented management, continuous improvement, involving the entire organization and
using data for decision-making, accounting firms can optimize processes, increase client
satisfaction and comply with legal regulations. These principles constitute a solid
framework for ensuring a constant level of performance and adapting to the constantly
changing requirements of the market.
Thus, a number of solutions can be considered: implementing a continuous training system
for employees (training sessions and workshops, as well as demonstrations of their benefits
for improving the quality and efficiency of accounting processes); adopting standardized
internal procedures or implementing an ERP (Enterprise Resource Planning) system; small
companies can opt for more accessible solutions, such as cloud accounting software, which
can automate many processes, or for the gradual implementation of a quality management
system, focusing on the most critical aspects; implementing regular feedback systems from
customers, assessments of their satisfaction and measurements of the efficiency of
accounting processes, such as response time and error rate in reports; adopting advanced
technological data security solutions (e.g. data encryption) and ensuring a clear internal
policy on the protection of confidential information; creating customer segments and
offering personalized service packages that meet specific needs, but which also comply
with the company's internal quality standards.

References
[1] Joseph Moses Juran, Planificarea calităţii, Editura Teora, Bucureşti, ISBN 973-601-
962-4, p.19, 2000
[2] Philip Kotler, Managementul Marketingului, Ed. Teora, București, ISBN 601-365-0 p.
585 – 587, 1999
[3] Andreea-Elisabeta Budacia, Lucian C-tin Gabriel Budacia, Is the Accounting of the
Future online?, Romanian Economic Business Review, Fall 2020, Volume 15, Number 5,
ISSN 1842 – 2497, p. 55 –60

Pag. 9 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[4] Lucian C-tin Gabriel Budacia, Bazele contabilității – elemente teoretice si practice,
Ed. Universitară, Bucureşti, ISBN 978-606-28-1868-5, p. 32, 2024
[5] https://www.iso.org - ISO 9001:2015 Standard of Quality management systems

Bibliography
E. A. Budacia, A. Perju – Mitran, Risk and Quality Management, Ed. Universitară & Ed.
Universul Academic, București, 2021
SRO – Catalogul Standardelor Române 2000, Ed. Tehnică, Bucureşti
Stanciu I. – Managementul calităţii totale – Ed. Cartea Universitară, Bucureşti, 2003
James R. Evans, William M.Lindsay – The management and control of quality, Mason
South-Western, 2008
Cristoph Loch, Stephen E. Chick, Arnd Huchzermeir – Managementul calităţii totale –
Ed. Cartea Universitară, Bucureşti, 2003
M.Lindsay – Management quality and competitiveness, Berlin Springer, 2008

Pag. 10 / 350
Article’s total number of pages: 10
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

BUILDING A SUSTAINABLE FUTURE BY DEVELOPING THE CONCEPT OF

DOUBLE DIAMOND IN TRIPLE DIAMOND, A NEW CONCEPT IN CREATING
A NUDGE TOWARDS PRO-ELECTRIFICATION BEHAVIOR IN URBAN
FREIGHT TRANSPORT

Irina CALOTĂ1
Alexandra PERJU-MITRAN2
Augustin SEMENESCU3

Abstract
In the present economic setting, road freight transportation accounts for 77% of the total
products moved by land inside the European Union. However, it is undeniable that, from a
certain perspective Historically, vehicles were detrimental to the environment. The freight
transport sector accounts for one-fourth of greenhouse gas emissions from road transport
and roughly 6% of overall greenhouse gas emissions in Europe. Consequently, the vehicle
electrical industry is poised for growth in Europe in the future.
Electrification is a growing trend driven by the necessity to reduce carbon emissions,
enhance air quality, and ensure compliance. Regulations governing urban emissions are
stringent.
These policies are contextualized within the European Union's target of achieving "a
reduction of at least 55% of greenhouse gas emissions by 2030 and climate neutrality by
2050," as outlined in the European Green Deal.
The Truck E-mobility concept in urban agglomerations pertains to the use of electric
vehicles, both high and small tonnage, for freight transportation in urban settings. This
movement is driven by the necessity to mitigate pollution, enhance logistical efficiency,
and comply with stringent regulations in densely populated urban areas.
To promote sustainable development, we propose the saddle transformer within the
framework of a novel Diamond idea based on the Double Diamond model. The first offers
a strategic design framework applicable to innovation processes and issue-solving. The

1 PhD Candidate, National University of Science and Technology Politehnica Bucharest, Bucharest, Romania
2 PhD, Romanian-American University, Bucharest, Romania, alexandra.perju-mitran@rau.ro, corresponding
author
3 PhD, National University of Science and Technology Politehnica Bucharest, Bucharest, Romania, Academy

of Romanian Scientists, augustin.semenescu@upb.ro

Pag. 11 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Triple Diamond is a novel idea that effectively provides a strategic framework for
implementing solutions in problem resolution.
Keywords: Sustainable Transport, Green Economy, Truck E-mobility, Double Diamond,
Triple Diamond, Nudging
JEL Classification: Q50, Q55, Q56

1. Introduction
The concept of a green economy, first introduced in 1989 in a report for the United Kingdom
government by a group of environmental economists titled "Blueprint for a Green
Economy" [1], fundamentally underpinned the implementation of an inclusive green
savings strategy at the United Nations Conference on Sustainable Development in 2012
(Rio+20) [2].
The definition provided by UNEP (2011) [3] is one of the most recognized internationally
and widely utilized, stating that a "Green Economy" is an economy that enhances the
welfare of populations and promotes social equity, while simultaneously significantly
reducing environmental risks and constraints. It is an economy characterized by minimal
carbon emissions, resource efficiency, and social inclusivity.
The notion of a green economy has garnered substantial international attention in recent
years, both as a mechanism for resolving the financial crisis of 2008 and as one of the two
themes for the United Nations Conference on Sustainable Development since 2012
(Rio+20). Consequently, this contributed to the growth of the concept through expanding
literature and rising international practices. It is essential to acknowledge that when the
notion was initially embraced as a theme for Rio+20, there was also ambiguity in EAEC.
What is the link between a green economy and internationally accepted objectives, such as
sustainable development and poverty eradication, as well as the lack of knowledge in the
EAEC? What are the possible problems, risks, costs, and rewards associated with the
adoption of a green savings policy?
The primary aim of the EAEC The envisioned economy is one that fosters a sustainable
future characterized by low carbon dioxide (CO2) emissions, achieved through a transition
to ecological production and consumption models that prioritize resource efficiency. This
work may be accomplished through the utilization of renewable energy, the adoption of
energy-efficient technology, and the implementation of circular economic models that
prioritize waste reduction and resource efficiency.

Pag. 12 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The shift to a green economy also has enormous hurdles, including the necessity for
substantial investment in technology and infrastructure, as well as the need to balance
economic, social, and environmental factors.

Road freight transportation accounts for 77% of the total products moved by land inside the
European Union and produces a commercial surplus of billions of euros yearly for the EU.
They are a significant and integral element of the economy. The freight transport sector
accounts for one-quarter of greenhouse gas emissions from road transport and roughly 6%
of overall greenhouse gas emissions in Europe. Consequently, the vehicle electrical market
will expand in Europe in the future.
Considering the present setting in which commodities are predominantly carried via road,
a mode responsible for 72% of total CO2 emissions from the transport sector, global efforts
are being made to promote electric freight transport. These policies are contextualized
within the EU's target of lowering greenhouse gas emissions by at least 55% by 2030 and
achieving climate neutrality by 2050, as outlined in the "European Green Deal."
Governments, corporations, and individuals must collaborate to address these challenges
and expedite the transition to electrified urban freight transport.
Electric mobility for trucks (truck e-mobility) plays a vital part in the transition to more
sustainable and environmentally sound transportation in urban environments.
Truck E-mobility [4] in urban agglomerations pertains to the utilization of electric vehicles,
both big and small tonnage, for freight transportation inside urban environments. This
movement is propelled by the necessity to diminish pollution, enhance logistical efficiency,
and comply with stringent regulations in densely populated urban areas.
What is the significance of implementing electrification of vehicles in urban areas?
• Air pollution reduction: electric trucks eliminate greenhouse gas emissions (CO₂) and
atmospheric pollutants (NOx, fine particulate matter).
• Strict regulations about emissions: several European cities, including London and Paris,
are instituting Low Emission Zones (LEZ), which restrict diesel cars, while also mandating
a shift to zero-emission vehicles.
• Reduction of urban noise: electric trucks are considerably quieter than their diesel
counterparts. What factors help to the reduction of pollutant phonics?
• Delivery request Rapid and sustainable: rise in e-commerce and anticipations client for
delivery. The advent of rapid transit has resulted in a heightened influx of automobiles in
urban areas; thus, electric trucks are a viable answer for more sustainable transportation.
Electric Mobility for Trucks Urban challenges:

Pag. 13 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Autonomy of batteries: electric trucks and vehicles possess a limited range, making them
more ideal for short-distance transportation or local delivery rather than intercity transit.
• Charging infrastructure: the scarcity of charging outlets in urban areas and logistics
warehouses is a significant hurdle.
Initial expenditures for electric trucks are greater than those for diesel vehicles, however
their operational costs are comparatively lower.
• Pregnancy is advantageous: batteries are bulky and heavy, reducing vehicle load capacity.
Integration with electric networks: the rising number of electric cars may need increased
energy requirements for networks, which must be equipped to manage rapid loading well.
Emerging solutions and trends in Truck E-mobility:
• Dedicated charging infrastructure: ultra-fast charging stations for trucks, strategically
located in urban areas and near logistics depots; logistics hubs equipped with renewable
energy sources (e.g., solar panels) for charging purposes.
• Advanced battery technology: high-density batteries provide more power capacity,
potentially increasing autonomy and reducing charging times.
• Pilot projects in major cities: firms such as Amazon, DHL, and IKEA are testing electric
truck fleets for urban delivery, while manufacturers like Tesla Semi, Volvo FL Electric, and
Mercedes eActros are introducing electric trucks for urban transportation.
• Innovations in logistics: last-mile deliveries facilitated by drones or tiny electric vehicles
(cargo bikes) and the exchange of data between enterprises for route optimization and a
decrease in the number of cars on the streets.
• Governmental subsidies and policies: governments provide incentives. Funding for the
procurement of vehicle electrical systems and the construction of charging infrastructure.
Examples of effective practices with the adoption of Truck E-mobility:
• London: Ultra-Low Emission Zones (ULEZ) [5] promote the transition to electric cars
for cargo delivery and electric truck fleets utilized by maritime merchants.
• Amsterdam: "Zero Emission Zones" Plan [6] (until 2025) which restricts entry for diesel
cars in some areas of the city and establishes Micro-Logistics Hubs for efficient deliveries
using electric vehicles.
• Copenhagen: Utilize electric cars for all government supplies and increase the network of
charging stations.
Long-term benefits of Truck E-mobility in urban areas:
• Operational cost reduction: Electric energy expenses are lower than those for fossil fuels,
and maintenance is very straightforward.
Pag. 14 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Sustainability and positive image: Companies that employ electric vehicles may enhance
their reputation, appealing to environmentally conscious clients.
• Growth Efficiency: Electric cars may be combined with autonomous driving technology,
significantly enhancing logistics efficiency.
Truck E-mobility is essential for addressing contemporary urban difficulties; yet, its success
relies on collaboration among manufacturers, governments, and logistics operators.
Infrastructure adaptation and cost-reduction technologies will expedite the transition to
electric transportation in densely populated cities.
Truck E-mobility is a crucial aspect of the transition towards sustainable development and
a sustainable future, linked with global aims to combat climate change and mitigate the
negative environmental impacts of transportation. Electric mobility in the vehicle sector
provides answers for many economic, social, and environmental concerns, significantly
contributing to progress. Sustainable electric mobility for trucks (truck e-mobility) plays a
vital part in the transition to more sustainable and ecological transportation.
Consequently, while sustainability and sustainable development are interconnected ideas,
they possess unique meanings.
Sustainability denotes the capacity of a system or process to be sustained at a specific level
across time. In a background setting, it signifies the necessity of generating current
meetings without jeopardizing future capacity production while fulfilling one's own
requirements. Sustainable development is a comprehensive term that encompasses not only
environmental sustainability but also economic and social sustainability.
The most recognized definition of sustainable development is provided by the World
Commission on Environment and Development (WCED) in the Brundtland Report:
"Sustainable development is development that meets the needs of the present without
compromising the ability of future generations to meet their own needs." [7]

2. Nudging
Nudging is a concept from behavioral sciences, economics, and psychology that focuses on
subtly influencing individuals' actions without constraining their choices or significantly
altering economic incentives. It is a method to influence behavior towards a desired
outcome by modifying underlying decision-making processes.
The notion of nudge encompasses several strategies employed to assist individuals in
making superior decisions without imposing certain outcomes on anybody.
Nudging was introduced in 2008 by Richard Thaler and Cass Sunstein and is described as
"any aspect of the choice architecture that modifies individuals' behavior in a predictable
manner, without prohibiting any options or significantly altering economic incentives." [8]
Pag. 15 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

For an intervention to qualify as a simple motivation, it must be easily avoidable and

inexpensive.
The word was popularized by economists Richard Thaler and Cass Sunstein in their book
"Nudge: Improving Decisions About Health, Wealth, and Happiness." [9] In the context of
time, "architecture of choice," as defined, pertains to "the environment" in which humans
make decisions. [10]
The strategist begins with the premise that sustainable development offers a conceptual
framework that, when embraced by citizens, will facilitate the establishment of a fair
enterprise characterized by balance and solidarity, enabling adaptation to changes induced
by global, regional, and national challenges, including demographic decline.
Skepticism and criticism are common reactions in the rest of the world about "nudges," due
to the fact that their objective is to alter an individual's behavior. Some authors regard
nudging as both protective and patronizing.
Attributes of nudging:
• Subtlety: Does not enforce rules or prohibitions, but facilitates the decision-making
process to promote outcomes that are more beneficial for the person or society.
• Freedom of choice: Does not restrict available options.
• Influence of context: Decisions are affected by the mode in which options are presented.
Advantages: Assists individuals in making judgments that may be more advantageous or
prudent in the long run; Costs are minimal compared to established laws.
Criticism may be perceived as manipulative, particularly if goals lack transparency; its
effectiveness is contingent upon cultural context and the manner of application.
Nudging is a strategy of persuasion that maintains individual individuality while subtly
encouraging desirable behaviors, particularly in the context of enhancing urban freight
transit in Romania alongside European Union obligations.
Punctuality in the context of assimilating electrification in urban freight transport arises
when involved parties encounter challenging and unique scenarios that complicate rational
decision-making. In such instances, ISI proposes a well-structured nudge or incentive:
Instances of nudging in the integration of Truck E-mobility inside urban freight transport
in Romania (source, author):
• Online environment: Brand awareness through viral video content produced by an
influencer on platforms such as Facebook and Instagram, focusing on user engagement;
Social messaging: Collantare Trucking's electric vehicles promoting "pro-
electrification/pro-neutrality" messages regarding climate and CO2 reduction, aimed at
raising awareness among all interested parties.

Pag. 16 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3. Double diamond
The Double Diamond is a design process concept popularized by the British Design Council
in 2005. [11] The technique was modified following the divergence-convergence model
introduced in 1996 by the Hungarian - American linguist Béla H. Bánáthy. [12] [13]. It is
prominent in user-centric design because it gives a clear direction for an approach confronts
complexity in a systematic and innovative manner.
The Double Diamond framework encompasses two diamonds that integrate four design
phases, showing a comprehensive overview of the design process.
1. Discoverr • Purpose: Comprehending context, identifying user needs, and recognizing
difficulties. • Methods: User interviews; Observation; Secondary research (data analysis,
existing studies) Outcome: A definitive understanding of the actual situation and a
compilation of insights.
2. Define • Purpose: Concentrate on the primary issue to be addressed, encapsulating the
insights gathered during the Discovery phase. • Methods: Formulation of issue statements;
development of empathy maps; mapping of customer journeys • Outcome: A clearly
articulated and well-defined design challenge.
3. Development • Objective: To generate and evaluate potential solutions for the identified
problem. • Methods: Brainstorming; Prototyping; Rapid Testing (User Testing) Outcome:
A collection of prototypes or concepts that address the issue.
4. Delivery • Purpose: Execute the completed solution and initiate its deployment. •
Methods: Completion Design; Implementation Solution; Measurement Impact • Outcome:
A functioning solution, prepared for implementation.
The visual representation (Figure No. 1) of the “Double Diamond” idea delineates two
divergent processes (exploring several choices) and two convergent processes (focused on
potentially viable solutions):
1. The initial diamond (problem): Divergence for exploratory issues and convergence for
their definitions; 2. The subsequent diamond (solution): Divergence for exploring solutions
and convergence for the implementation of optimal solutions.
Benefits of Double Diamond framework:
• Clarity and Structure: Offers a systematic approach to resolving complicated challenges.
• User-centric: Place people and their needs at the core of the process.
• Flexibility: Applicable across many domains, including product design, services, and
governmental policies.

Pag. 17 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Collaboration and Creativity: Promotes the engagement of interdisciplinary teams and

rapid experimentation.

Critique and constraint:

• May be regarded as linear, while the process is, in reality, iterative.
• Requires time and resources for efficient implementation. Success is contingent upon the
quality of research and user engagement.
3.1. Practical Applications of Double Diamond Product Design: Development of solutions
that directly address consumer demands. • Public Policies: Enhanced and more efficient
policy execution. • Digital Services: User-centric application development and platform
services.
The Double Diamond framework assists teams systematically addressing difficult
challenges while keeping the user at the center of the process.

Figure 1. Double Diamond / Design Council of the United Kingdom in year 2019
Source: https://projekter.aau.dk/projekter/files/415128891/

4. The connection between Nudging and Double Diamond

Pag. 18 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Nudging and the Double Diamond are complimentary ideas applicable in user-centric
solution design, particularly in service design, public policy, and consumer behavior.
Despite originating from disparate contexts, they can be interrelated to inform the
development of successful treatments that impact human behavior. What examines the
architectural modeling choices about the use of electricity in urban freight transport and
promotes carriers' selection of sustainable options.
Nudging may be included in the Double Diamond approach to provide solutions that subtly
impact habits while preserving user autonomy.
1. Discover:
• Identify behavioral streams and the elements that drive them.
• Investigate users to comprehend the rationale behind my decisions (cognitive biases,
habits, obstacles).
Examples: Observing decisions made by carriers to comprehend my continued preference
for fossil fuel trucks over electric ones.
2. Define:
• Formulating the problem: "How can we devise a solution that promotes desired
behavior?"
• Problem statements may encompass: “What strategies may be employed to enhance the
electrification rate among carriers?" How can we motivate carriers to conserve energy
through saddle SAVE?
• Conduct data analysis to determine the locations and methods for implementing a nudge
for A.
3. Formulate:
• Formulating innovative solutions grounded in the principles of nudging, including:
• Default settings: Implicitly configuring certain preferred options (e.g., automatic
enrollment in savings programs).
• Social Messages: Communication FACT that "countries such as Germany, France, and
the Netherlands have integrated electrification in urban freight transport" to promote
adherence.
• Redesign of environmental decision-making: Mandate by legislation.
• Prototyping and evaluating solutions through simulation or pilot studies.
4. Delivery:
• Execution of solutions and assessment of impact.

Pag. 19 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Examples: Organizing pilot truck "drive test" sessions for electric vehicles by
manufacturing for carriers in major metropolitan agglomerations.
• Modify energy bills for A to emphasize consumption in comparison to other nations that
have already included electric trucks into urban transport.
In what ways does nudging facilitate the Double Diamond process?
• Validating insights: Researching behavior particular to nudging (observing decisions,
testing reactions) might enhance the Discovery phase.
• Prototypes are expedited and iterative: During the development phase, nudging provides
straightforward and efficient solutions that can be rapidly evaluated.
• Measurement of success: Nudging relies on minor, quantifiable modifications, facilitating
an easier evaluation of impact during the delivery phase.
Advantages of integrating Nudging with the Double Diamond framework:
• Resolutions User-centric: The Double Diamond framework directs the creative process,
while nudging guarantees that the solutions are pragmatic and effective.
• Scalability: Nudging facilitates widespread deployment while minimizing expenses.
Quantifiable impact: The integration of user research (Double Diamond) and behavioral
testing (nudging) enhances the likelihood of success.
In summary, Double Diamond provides a framework for problem-solving, whereas
Nudging introduces a behavioral dimension that enhances the offered solutions. This
combination is highly beneficial in formulating effective and lasting solutions.

5. Outcomes
To achieve sustainable development in Romania in alignment with European Union
commitments, we must formulate a strategy centered on citizens and future generations
concerning the implementation of global trends in "Truck E-mobility" through
complementary concepts: NUDGING and DOUBLE DIAMOND.
Specifically, in the context of AUTHOR development, the transformation of the Double
Diamond model into a Triple Diamond framework, once embraced by all stakeholders in
Truck E-mobility, will facilitate the establishment of a company characterized by equity
and solidarity, capable of addressing the challenges posed by global, regional, and national
changes.
In contrast to the Double Diamond, which offers a strategic design framework applicable
to innovation processes and issue-solving, the Triple Diamond effectively provides a
strategic framework for the execution of solutions in problem resolution.
Pag. 20 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

examined the "Double Diamond" idea inside EAEC. The applicability of electrification in
urban freight transport is examined by the author, who notes that the two components—
Discovery/Definition and Development/Delivery—are intricately connected. However,
they must be addressed in a manner that facilitates both a visual representation of standard
design phases (problem/solution) and an overview of the implementation/testing process.
This article necessitates the development of three diamonds, envisioned as the "Triple
Diamond."
Third diamond visible The author possesses the subsequent phases:
- Training Driver (divergence) / Test Drive (convergence), denoting the procedures
designed to apply and evaluate the solution derived from information acquired through
observation of the necessity to facilitate the adoption of electric vehicles in urban freight
transport.

Figure 2. Triple Diamond: Representation visual modified Double Diamond model

encapsulates the design process How and implementation
Source: author / contributions PERSONAL

In the Triple Diamond model, the author illustrates that the three diamonds symbolize a
process of researching challenges, which may be extensive or profound (divergent
thinking), executing targeted actions (convergent thinking), and then implementing
solutions. Therefore, the design process has six processes, exceeding the two additional
phases typically illustrated in literature according to the Double Diamond design process
model, which was popularized by the British Design Council in 2005, specifically:

Pag. 21 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Discover: comprehend the situation rather than make assumptions about what it is. This
step involves engaging in discussions and socializing with individuals impacted by issues.
Define: with the information gathered from the discovery phase, define CHALLENGE in
a different way. The concept of challenge utilizing the insights acquired during the
discovery phase.
Develop: provide answers different to the problem definition clearly, looking for inspiration
from elsewhere part and co -designing with several people different.
Delivery: offer different solutions at scale small.
Training drivers: offer dedicated training sessions for different truck models electric
vehicles intended for urban use.
Test Drive: evaluate several electric truck models designed for urban applications for a
duration ranging from 2 to 7 days or a minimum distance of 100 km driven, referred to as
"Truck E-mobility".
The two additional stages facilitate the monitoring of views and attitudes toward the
organization, its goods and services, as well as customer satisfaction and perceptions of
mark, and are likely to provide challenges. Including analysis, the candidate is gathering
feedback on the company's exploitation by conducting test drives over distances not
exceeding 100 km, assessing performance and reception in comparison to rivals.
Through the analysis of this data by the marketing and sales departments, along with other
project team members, market gaps can be identified, as well as opportunities for
enhancement and innovation in the assimilation and implementation of the "Truck E-
mobility" concept from regulatory objectives to practical application.

6. Conclusion
Under the circumstances of "Truck E-mobility," cities face several challenges across
various sectors, including the economics, energy, transportation, infrastructure, water
supply, environmental protection, and essential services.
To achieve the objectives set by the European Union to reduce greenhouse gas emissions
by at least 55% by 2030 and attain climate neutrality by 2050, as outlined in the "European
Green Deal," Romania must now devise solutions for the assimilation and implementation
of "Truck E-mobility," thereby transforming into one of the future's intelligent cities.
The research indicated that exploring and comprehending fundamental components from
specialized literature concerning the brain's organizational modules and their differential
responses to events is essential for creating a "nudge" to facilitate the assimilation and
implementation of the phenomenon of "Truck E-mobility."

Pag. 22 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The literature on Nudging and the Double Diamond framework has influenced our research
in the creation of the Triple Diamond idea, which centers on the end user of the electric
vehicle and may be applied to saddle design. Offer results good outcomes essential for the
decarbonization of urban freight transport and line with EU targets (years 2030/2050).
Addressing the mentioned problem in a sustainable manner may involve the adoption of
electric truck growth, as evidenced by considerable literature study. We provided a
comprehensive analysis of the variables affecting the delayed adoption of electric cars and
the constraints that we, as service designers, must consider. When Cream resolves our
conceptual issues.
By examining and evaluating recommendations from the specialist literature to develop an
effective nudge, the author proposes utilizing these insights in conjunction with our service
design tools and the methodologies at our disposal. A research question may be worded as
follows: “How can service design be utilized to address the needs of individuals based on
sustainable lifestyles?” This research topic will lead us in designing actual, relevant
solutions through a case study.
We recognized some notable strengths while also emphasizing areas for future enhancement
across a variety of literature, including foundational books on the green economy and
sustainable development, as well as specific frameworks such as nudging and the Double
Diamond model. This scope provides a robust theoretical framework and contextualizes
the importance of truck e-mobility within the broader framework of urban sustainability.
The author adeptly integrates design strategy with execution by extending the Double
Diamond concept into a Triple Diamond framework.
This conceptual development redefines the design process and underscores the practical
challenges of executing sustainable transportation solutions. Our research sought to address
the urgent need to reduce emissions in urban transportation while adhering to EU policy
goals. The focus on truck electrification due to environmental and regulatory requirements
is entirely justified. The discussion on nudging presents an intriguing behavioral dimension
to the findings. The essay outlines a sophisticated approach to accelerating the slow
adoption of electric vehicles by analyzing how small interventions might affect decision-
making. While the theoretical framework is robust, our study heavily depends on existing
literature without including significant empirical data or case studies.
Future research may benefit from pilot studies or field trials that validate the proposed
Triple Diamond structure in real-world settings.
An additional limitation of our current research is that the transition from the established
Double Diamond to the proposed Triple Diamond may be articulated with better precision.
Detailed explanations of each phase, especially the "Training Driver" and "Test Drive"
components, would improve readers' comprehension of the framework's implementation.
Despite the conceptual model's novelty, there is inadequate discussion of the practical
Pag. 23 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

challenges and limitations (e.g., infrastructure, financial implications, stakeholder

participation) that may affect its adoption. A thorough evaluation of these challenges would
improve the analysis. Moreover, using quantitative metrics (such as performance statistics,
cost-benefit analyses, or impact assessments) might provide a more balanced view between
theory and practice. Future initiatives may include pilot projects or case studies employing
the Triple Diamond framework in actual urban transportation settings. This would not only
validate the model but also highlight any necessary improvements for different
circumstances. Furthermore, given the multidisciplinary nature of the issue, a thorough
stakeholder analysis might provide insights into how diverse organizations (governments,
logistics companies, urban planners) can collaborate to resolve implementation challenges.
Investigating the lasting impacts of nudging strategies on behavior and the effectiveness of
electric vehicles in urban logistics would yield valuable insights into the sustainability and
scalability of this method.

References
[1] UNDESA (2012) - A guidebook to Green Economy (2012), Issue 1: Green Economy,
Green Growth, and Low Carbon Development – history, definitions and a guide to recent
publications, page 5, 7 [26.11.2024]. Available:
https://sustainabledevelopment.un.org/content/documents/GE%20Guidebook.pdf
[2] UNDESA (2012), A guidebook to Green Economy (2012), Issue 2: exploring green
economy principles, p.3 [26.11.2024]. Available:
https://sustainabledevelopment.un.org/content/documents/743GE%20Issue%20nr%202.p
df
[3] UNEP, (2011) Towards a Green Economy: Pathways to Sustainable Development and
Poverty Eradication, 2011, p. 9 [26.11.2024].
[4] Electric Mobility – Wikiversity [26.11.2024] Available:
https://en.wikiversity.org/wiki/Electric_Mobility
[5] Wikipedia – Ultra Low Emission Zone [26.11.2024]. Available:
https://en.wikipedia.org/wiki/Ultra_Low_Emission_Zone
[6] Wikipedia – Low-Emission Zone: [26.11.2024]. Available:
https://en.wikipedia.org/wiki/Low-emission_zone
[7] Our Common Future: The Report of the World Commission on Environment and
Development. Oxford University Press, Oxford & New York: In Oxford Paperback, 1987.
[26.11.2024]. Available: www.un-documents.net/our-common-future.pdf

Pag. 24 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[8] Thaler RH, Sunstein CR, & Balz, JP (2012). Choice architecture. Behavioral
Foundations of Public Policy. [26.11.2024]. Available:
https://doi.org/10.2139/ssrn.1583509
[9] Thaler RH, Sunstein CR (2008), Nudge: Improving decisions about health, wealth,
and happiness: Yale University Press. [26.11.2024]
[10] Design Council. (2019b, October). The Double Diamond: A universally accepted
depiction of the design process. https://www.designcouncil.org.uk/news-opinion/double-
diamond-universally-accepted-depiction-design-process . [26.11.2024]
[11] Design Council, Eleven lessons. A study of design process, 2 December 2022, p 6-7.
[26.11.2024].
[12] Goldstein DG, Johnson EJ, Herrmann A., Heitmann M. (2008). Nudge your
customers towards better choices, Harvard Business Review, vol. 86, pp. 99-105
[13] Selinger E., Whyte K. (2011). Is there a right way to nudge? The practice and ethics
of choice architecture, Sociology Compass, vol. 5, pp. 923-935. [26.11.2024]
[14] Sunstein C. (2015). The Ethics of Nudging. In: Yale Journal on Regulation 32, pp.
413–450. [26.11.2024].

Bibliography
Carolin R., George B., Toby P., Neale K. (2020). Driving and accelerating the adoption of
electric vehicles in the UK, 239. Retrieved on 15.11.2024.
Cass R. Sunstein, Richard Thaler, Richard H. Thaler (2008). Improving Decisions About
Health , Wealth , and Happiness , pp.67-68. Retrieved 15.11.2024.
Caulliraux , AA, Bastos, DP, Araujo , R., & Costa, SR (2020). Organizational
optimization through the double diamond-Applying Interdisciplinarity . Brazilian journal
of operations & production management, 17(4), pp. 1-12.
Design Council, Eleven lessons. A study of design process, 2 December 2022, p 6-7.
[26.11.2024].
Design Council. (2019b, October). The Double Diamond: A universally accepted
depiction of the design process. https://www.designcouncil.org.uk/news-opinion/double-
diamond-universally-accepted-depiction-design-process. [26.11.2024]
Electric Mobility – Wikiversity [26.11.2024] Available:
https://en.wikiversity.org/wiki/Electric_Mobility
Goldstein DG, Johnson EJ, Herrmann A., Heitmann M. (2008). Nudge your customers
towards better choices, Harvard Business Review, vol. 86, pp. 99-105
Pag. 25 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Our Common Future: The Report of the World Commission on Environment and
Development. Oxford University Press, Oxford & New York: In Oxford Paperback, 1987.
[26.11.2024]. Available: www.un-documents.net/our-common-future.pdf
Selinger E., Whyte K. (2011). Is there a right way to nudge? The practice and ethics of
choice architecture, Sociology Compass, vol. 5, pp. 923-935. [26.11.2024]
Selinger E., Whyte K. (2011). Is there a right way to nudge? The practice and ethics of
choice architecture, Sociology Compass, vol. 5, pp. 923-935. Retrieved on 27.11.2024.
Sunstein C. (2015). The Ethics of Nudging. In: Yale Journal on Regulation 32, pp. 413–
450. [26.11.2024].
Thaler RH, Sunstein CR (2008), Nudge: Improving decisions about health, wealth, and
happiness: Yale University Press. [26.11.2024]
Thaler RH, Sunstein CR, & Balz, JP (2012). Choice architecture. Behavioral Foundations
of Public Policy. [26.11.2024]. Available: https://doi.org/10.2139/ssrn.1583509
UNDESA (2012) - A guidebook to Green Economy (2012), Issue 1: Green Economy,
Green Growth, and Low Carbon Development – history, definitions and a guide to recent
publications, page 5, 7 [26.11.2024]. Available:
https://sustainabledevelopment.un.org/content/documents/GE%20Guidebook.pdf
UNDESA (2012), A guidebook to Green Economy (2012), Issue 2: exploring green
economy principles, p.3 [26.11.2024]. Available:
https://sustainabledevelopment.un.org/content/documents/743GE%20Issue%20nr%202.p
df
UNEP, (2011) Towards a Green Economy: Pathways to Sustainable Development and
Poverty Eradication, 2011, p. 9 [26.11.2024].
Wikipedia – Low-Emission Zone: [26.11.2024]. Available:
https://en.wikipedia.org/wiki/Low-emission_zone
Wikipedia – Ultra Low Emission Zone [26.11.2024]. Available:
https://en.wikipedia.org/wiki/Ultra_Low_Emission_Zone

Pag. 26 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

RETHINKING HR MANAGEMENT FOR THE DIGITAL WORKPLACE

Niculina CHIVU 1

George CARUTASU 2

Abstract
Building on recent theoretical advances, including digital human capital theory,
organizational adaptability, strategic alignment, and technology adoption, this study
examines how digital HR tools improve organizational agility, personalization, and
decision making while maintaining a focus on people-centric strategies. A two-tiered
evaluation framework was proposed to assess both organizational outcomes and employee
perceptions, enabling a balanced analysis of hybrid HRM practices. This framework
supports evidence-based decision-making by combining quantitative performance
indicators with qualitative feedback mechanisms. Supported by industry examples and
benchmarking data, this study highlights the benefits of aligning digital solutions with
strategic HR objectives, while acknowledging implementation challenges related to data
security, organizational readiness, and employee adaptation. Additionally, this study
highlights the role of predictive analytics and continuous learning in optimizing HR
interventions. The findings reinforce the need for gradual, evidence-based adoption of HR
technologies and provide practical guidance for designing agile, inclusive, and sustainable
HRM systems that are ready for the complexities of the digital age.
Keywords: Human resource management, Digital transformation, Strategic alignment,
Employee engagement, Performance appraisal, Predictive analytics, Hybrid HR strategies
JEL Classification: M12; M53; D24

1. Introduction

Human resource management is of great importance in organizations, as it deals with the

recruitment, selection, training, development, appraisal, evaluation, and reward of
employees. The main aim is to provide the human resources needed by the organization to

1 PhD Candidate, National University of Science and Technology Politehnica Bucharest, Romania,
niculina.chivu@stud.fiir, corresponding author
2 Prof. PhD. Hab. Romanian-American University, National University of Science and Technology

Politehnica Bucharest, Romania, george.carutasu@rau.ro

Pag. 27 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

achieve its strategic objectives. Human resource management aims to optimize the
performance of human resources by creating a favorable and beneficial work climate and
developing employees. A well-trained, motivated, and performing team can differentiate
between the success and failure of an organization.

A central theme emerging from contemporary research is the concept of digital human
capital, which expands the traditional understanding of human capital by including digital
skills and adaptability. According to digital human capital theory [1], organizations need to
invest not only in attracting talent with digital capabilities but also in nurturing and retaining
talent through continuous learning opportunities and a culture of innovation. In the context
of a competitive digital labor market, the ability to build and sustain digital competencies
internally becomes a strategic advantage.

Adaptability is another pressing challenge for organizations. As external environments

evolve rapidly, there is a growing need for employees to continually update their skills and
respond effectively to organizational changes. The theory of organizational adaptability
[13], posits that employees who cultivate adaptability skills are better equipped to integrate
new technologies and workflows. HRM strategies that support continuous learning,
flexibility, and psychological safety create conditions for such adaptability.

The strategic alignment between employee needs and corporate goals remains an enduring
concern in HRM. The theory of strategic alignment in HRM [7], emphasizes the need to
synchronize HR initiatives with the overall goals of the organization. When performance
management, career development, and reward systems are aligned with both strategic
priorities and individual aspirations, organizations benefit from increased employee
engagement and more consistent performance outcomes.

In addition, promoting diversity and inclusion is increasingly being recognized as both a

social and a performance driver. Organizational inclusion theory [15], emphasizes the value
of creating work environments in which diverse employees feel respected, represented, and
empowered. Inclusive human resource management practices such as bias-aware
recruitment, promotion equity, and inclusion training contribute to greater innovation,
collaboration, and employee well-being.

High staff turnover, another major challenge, is often linked to poor work-life balance. The
work-life balance theory [6], argues that employee satisfaction and retention improve when
individuals are supported in managing their professional and personal roles. HRM policies
that include flexible work schedules, remote work options, and personal development
support have been shown to significantly reduce attrition and improve organizational
commitment.

Leadership development also plays a critical role in organizational success. The

transformational leadership theory [2], asserts that leaders who inspire and empower their
Pag. 28 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

teams promote high levels of engagement and innovation. HRM interventions that focus on
cultivating transformational leadership competencies through training, coaching, and
feedback can have far-reaching effects on team dynamics and organizational outcomes.

Finally, the successful adoption of HR technologies is essential for increasing the efficiency
and transparency of processes. The theory of technology adoption in HRM [9], explores
how digital platforms can improve various HR processes from recruitment and onboarding
to performance appraisal and career planning. However, adoption depends on both the
technological readiness of the organization and the perceived usefulness of the tools by
employees and managers. Human resource management strategies must therefore address
not only system implementation but also change management, communication, and digital
literacy. Together, these theories provide a sound conceptual basis for designing HRM
systems that are strategic, responsive, and human-centered. In the digital age, efficient
human resource management is not only about automation, but also aims to create a
flexible, inclusive and technology-supported environment in which employees can evolve
and perform to their full potential.

2. The role of human resources management (HRM)

HRM is essential to the smooth functioning of an organization, with the main purpose of
ensuring a climate conducive to employee development and performance to achieve
organizational objectives. Human resource management focuses on overseeing human
capital by balancing organizational goals with employee rights and motivations.

Human Resource Management (HRM) holds a crucial position in influencing an

organization’s ability to reach its long-term strategic objectives. HRM contributes
significantly to improving performance, ensuring employee engagement, and maintaining
legal and ethical compliance through a combination of structural, developmental, and
relational functions. The following analysis presents the main strategic role of HRM and its
practical impact on modern organizations.

Recruitment and selection are core responsibilities of HRM. This function ensures that the
organization attracts and identifies candidates whose skills and competencies align with
specific job requirements and organizational values. For example, Google has partnered
with leading universities to access emerging talent through its internships. By identifying
and onboarding high-potential candidates before graduation, the company cultivates a
culture of innovation and alignment. This approach is underpinned by the human capital
theory, which emphasizes that investing in human resource education and development
yields significant long-term organizational benefits. The results of such strategies in Google
include sustained capacity for innovation and an agile talent pipeline.

Another key dimension of HRM is employee training and professional development (PD).
Ensuring access to continuous learning opportunities improves both individual
Pag. 29 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

competencies and overall organizational adaptability. Toyota exemplifies the best practices
in this area through its on-the-job training programs, whereby new employees acquire skills
directly in production environments alongside experienced mentors. This model mirrors
Kolb's experiential learning theory, which argues that learning is most effective when it
occurs through a direct experience. Consequently, Toyota was able to cultivate not only
technical competence but also a strong sense of team integration and loyalty.

Human resource management (HRM) plays a critical role in maintaining employee

motivation. Motivational strategies aim to sustain a positive organizational climate and
encourage long-term employee engagement. Spotify provides a compelling example of
adopting a flexible working model that allows employees to structure their own schedules,
including the option of remote working. These practices are rooted in the self-determination
theory, which suggests that autonomy enhances intrinsic motivation. The empirical results
at Spotify include higher levels of employee satisfaction and reduced absenteeism and
turnover rates, reinforcing the strategic value of flexibility.

Performance management is another strategic HRM function that involves monitoring and
evaluating employees’ contributions against predefined objectives. General Electric (GE)
has implemented a 360-degree feedback system, allowing for multidimensional evaluation
from supervisors, peers, subordinates, and, in some cases, customers. This comprehensive
approach is consistent with the multiple-source feedback theory, which posits that diverse
perspectives produce more balanced and objective performance evaluations. The practical
impacts of the GE system include improved diagnostic accuracy, more targeted
development plans, and improved performance alignment.

HRM also plays a key role in managing labor relations, including maintaining constructive
communication channels and resolving workplace conflicts. In Google, an internal
mediation mechanism facilitates early conflict resolution through confidential and non-
punitive processes conducted by specialized staff. This system aligns with the conflict
management theory, which emphasizes the value of structured dialogue in promoting
collaboration and maintaining a positive work environment. The introduction of such
mechanisms has contributed to a measurable reduction in workplace tensions and has
supported the development of cooperative organizational cultures.

Another fundamental responsibility of HRM is ensuring legal compliance. This involves

upholding labor standards, protecting employee rights, and adhering to evolving legislative
frameworks. Accenture provides an example of best practice through its dedicated HR and
legal units tasked with monitoring regulatory changes and implementing the necessary
adjustments. This commitment reflects legal compliance theory, which emphasizes the
strategic importance of regulatory adherence in minimizing legal risks and protecting
organizational legitimacy. The Accenture's strict compliance protocols have enabled it to
avoid litigation and maintain an environment of high trust among its workforce.
Pag. 30 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Together, these roles highlight the multidimensional nature of HRM as both an operational
and strategic function. By aligning recruitment, development, motivation, performance
management, conflict resolution, compliance with organizational goals, and theoretical
underpinnings, HRM serves as a central driver of sustainable performance and
organizational resilience in the digital age.

3. Functions of Human Resource Management

The responsibilities of Human Resource Management involve the set of actions carried out
by this department to support and fulfill the organization’s established goals. The primary
functions of HRM (presented in Table 1).

Human Resource Management (HRM) has a central role in organizations’ strategic and
operational success. Core HRM functions serve not only administrative goals but also
strategic transformation and long-term performance. Below, we explore the key HRM
functions by integrating relevant theories, practical examples, and observed results.

One of the fundamental functions of HRM is recruitment and selection, which aim to attract
and identify the most suitable candidates for organizational roles. For example, Google's
collaboration with top universities to find top talent through internship programs illustrates
a strategic recruitment approach. This aligns with human capital theory [1], which posits
that investment in skilled human capital generates long-term organizational benefits. As a
result, Google promotes an innovation-driven culture anchored in talent and aligned with
the company's vision.

Another key area is training and development, which ensure that employees continuously
evolve with the needs of the organization. Toyota provides an exemplary model through
hands-on training on the production line, allowing employees to learn by doing so. This
practice reflects Kolb's (1984) experiential learning theory, which emphasizes learning-by-
experience. This results in a skilled, engaged workforce with a deep sense of ownership and
connection to the company. The synthetic process is presented in Table 1.

HRM Description Practical Relevant Results

Function example theory

Recruitment Attracting the Google: Human Capital Innovative

and selection right University Theory culture, aligned
partnerships for with talents.

Pag. 31 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

candidates for the early search (Agarwal et al.,

open positions. for talent. 2020)

Training and Professional Toyota: On-the- Experiential Technical

development development job training at Learning growth and
improve the workplace Theory (Kolb, team
employee model. 1984) membership.
skills.

Performance Continuous GE: 360-degree Feedback Holistic

evaluation and objective evaluation theory from assessment and
performance system. multiple improvement.
monitoring. sources
(DeNisi &
Williams,
2020)

Motivation Maintaining Spotify: The Theory of Increased

and rewards satisfaction Flexible Self- satisfaction and
and loyalty. working model. Determination retention.
(Ryan & So,
2000)

Labor Creating a Google: Conflict Reduced

Relations harmonious Internal Management internal
Management working mediation Theory conflicts.
environment. system. (Rahim, 2002)

Strategic HR Long-term IBM: Digital Strategic Technology-

Planning planning transformation Human Capital based
aligned with and reskilling. Planning workforce
strategy. (Lawler & adaptation.
Boudreau,
2021)

Legal Compliance Google: GDPR Data Protection Legal risk

Compliance with labor compliance and Theory mitigation and
laws and data employee (Jarrahi, 2020) trust.
regulations. training.

Table 1 Principal functions of human resource management (HRM)

Pag. 32 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Performance evaluation plays a key role in the overall HR strategy. GE's use of 360-degree
feedback ensures a comprehensive review of each employee by peers, subordinates, and
managers. This approach is based on multi-source feedback theory [4], which promotes
objectivity and well-rounded evaluations. GE has improved individual performance by
providing balanced and actionable feedback tailored to individual developmental needs.

Motivation and reward systems significantly influence employee engagement. Spotify's

implementation of a flexible working model supports autonomy and work-life balance,
which are the basic tenets of self-determination theory [3]. This policy has led to increased
employee satisfaction and a notable decrease in turnover, proving that motivational
strategies directly affect employee retention.

Labor relations management is also vital for maintaining a positive organizational climate.
Google's internal conflict-mediation system exemplifies proactive relationship
management. Based on the conflict management theory [6], this approach emphasizes
addressing issues proactively and finding common ground, which helps foster better
teamwork and minimizes workplace friction.

Another strategic component of human resource management is human resource planning,

which forecasts future talent needs in line with business strategy. IBM exemplifies this
through reskilling initiatives as part of its digital transformation [10].

Google's GDPR compliance underscores the importance of data protection and regulatory
compliance. This is supported by the data protection theory [9], which emphasizes the role
of privacy and trust in HR operations. By training employees and enforcing strict privacy
protocols, Google mitigates legal risk and maintains stakeholder trust.

In short, effective HRM functions are based on a balance between theory-driven practices,
technology integration, and people-centered strategies. The synergy of these elements
improves organizational agility, innovation, and long-term performance.

4. Evolution of HRM methods: from classical approaches to digital solutions

Human resource management (HRM) has undergone a profound transformation, evolving

from traditional manual practices to digitally supported strategies designed to improve
efficiency, objectivity, and accessibility. This section compares the classic methods
historically used in HRM with contemporary digital approaches, highlighting the key
differences in recruitment, training, and performance appraisal.

Pag. 33 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

4.1. Classical methods in human resource management

Classical HRM methods are predominantly manual, time consuming, and often limited in
scope. These approaches were prevalent before the widespread adoption of digital tools,
and relied mainly on face-to-face interactions and paper documentation. In the field of
recruitment and selection, classic methods rely on advertising vacancies in newspapers,
attending job fairs, and using personal recommendations. The assessment of candidates
usually consisted of face-to-face interviews and handwritten assessments conducted
without the support of standardized instruments or data analysis. This approach, while
offering a degree of personalization, is resource intensive and often prone to subjectivity.

Training and development in the classical context involves face-to-face workshops and
mentoring sessions conducted on-site or in external training facilities. While these programs
facilitated personal interaction and the acquisition of practical skills, they required
substantial time and financial investment, limiting access and scalability. Employees had to
be physically present, which further restricted their flexibility and inclusion. Traditional
methods of performance appraisal typically involve annual appraisals based on written
reports and supervisor observations. Feedback is formal, periodic, and often delayed, which
reduces its effectiveness as a development tool. The subjective nature of these evaluations
raises challenges regarding consistency and fairness. While these methods laid the
foundation for professional HR practice, their limitations in scale, speed, and objectivity
created the need for more adaptive and data-driven approaches.

4.2. Digital methods in human resources management

With the advent of digital technologies, HRM has shifted to methods that offer improved
efficiency, broader reach, and data-driven decision making. Digital solutions have enabled
process automation, real-time performance monitoring, and personalization at this scale.

In recruitment and selection, digital platforms such as LinkedIn and (Applicant Tracking
Systems) have transformed the way organizations identify and assess talent. Artificial
intelligence is now being used to scan resumes, assess compatibility with job descriptions,
and support objective filtering. Video interviewing and automated assessments have further
accelerated the hiring process and expanded access to a global pool of candidates, thereby
reducing the time to hire and increasing diversity.

Employee training has changed, and e-learning platforms and learning management
systems (LMS) allow organizations to deliver customized training programs remotely,
enabling flexible and cost-effective upskilling. Courses can be tailored to individual needs,
and employee progress can be tracked and analyzed. This digital shift has democratized
access to high-quality learning resources and supports continuous development at all
employee levels.
Pag. 34 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

To assess performance, modern organizations are increasingly using software solutions that
enable continuous feedback, track KPIs in real-time, and generate comprehensive analytical
reports. These tools provide structured, measurable, and objective information about
employee performance, facilitating timely coaching and aligning individual contributions
with organizational goals. The continuous feedback culture fostered by these systems
increases transparency and supports continuous improvements.

A 2023 study by Undelucram (citation), involving over 5,000 employers and 200,000
employees, highlights the importance of digital HR tools in bridging the gap between
organizational offerings and employee expectations. Findings revealed that:

• 76.22% of employees prioritize bonuses and financial benefits;

• 31.85% prefer educational and development benefits;
• 94.24% of employers currently offer free or subsidized training programs;
• 90.65% offer meal vouchers.

These results highlight the discrepancies between the types of benefits offered and those
desired by employees, underscoring the need for the digital personalization of HR
strategies. By using digital tools, organizations can tailor benefit packages to individual
preferences, thereby increasing satisfaction, engagement, and retention.

In conclusion, while classic HRM methods have established essential frameworks for
personnel management, their limitations have been increasingly addressed through digital
innovation. The integration of digital technologies not only improves operational efficiency
but also enhances employee experience, supports strategic decision-making, and aligns HR
practices with contemporary organizational dynamics.

4.3. Comparison of classical and digital methods

The digital transformation of human resource management (HRM) has had a significant
impact on organizational processes, particularly in terms of operational expenses and
processing time. The implementation of digital technologies has enabled organizations to
streamline recruitment, training, performance appraisal, and employee data management
processes, resulting in substantial cost savings and increased efficiency.

Pag. 35 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 1 Comparison between classical and digital HRM methods

Aspect Classic methods Digital methods

Recruitment Newspapers, job fairs, Online platforms, application

recommendations management systems (ATS), video
interviews

Selection Traditional interviews, Automatic screening, AI for resume

manual tests analysis

Training Face-to-face courses, E-learning, LMS, interactive online

classic courses courses

Performance Written reports, annual Continuous feedback, performance

evaluation interviews appraisal software

Flexibility Limited High: access to trainings and

assessments anytime

Table 2. Multi-aspect comparison between classic and digital HRM methods

Pag. 36 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

A study conducted by [11], found that 83% of recruiters worldwide believe that recruitment
technology saves them significant time during the selection process. According to this
study, implementing an applicant tracking system (ATS) can reduce recruitment time by up
to 50% compared with traditional manual methods.

In addition, a 2022 report by consulting firm PwC indicates that automated recruitment
technologies can reduce the time it takes to go from advertising to hiring a candidate by at
least 30%.

5. Methodological framework for evaluating hybrid HRM strategies at organizational

and employee level

In the current landscape of accelerated digital transformation and increasingly dynamic

labor markets, human resource management (HRM) systems are challenged to evolve
beyond traditional frameworks. Contemporary organizations no longer operate on static
administrative processes alone but rather require a strategic, agile, and people-centric
approach that integrates digital technologies with conventional HR practices. In this
context, hybrid HRM models, which combine traditional methods with digital applications,
have emerged as an essential response to the increasing complexity and volatility of
organizational environments.

A hybrid HRM approach utilizes digital technologies such as applicant tracking systems
(ATS), e-learning platforms, performance dashboards, and predictive analytics tools while
retaining the relational and contextual richness of face-to-face interviews, coaching
sessions, team meetings, and interpersonal communication. This combination improves
recruitment efficiency, enables personalized learning trajectories, supports real-time
feedback in performance appraisal, and ensures a balanced relationship between employees
and the organization. However, the success of such an integration depends to a large extent
on the organization's ability to assess both the operational effectiveness and human impact
of these practices.

To address this need, we propose a methodological framework based on a two-level

evaluation matrix that simultaneously considers organizational performance indicators and
employee-level perceptions. This two-dimensional model enables a comprehensive
diagnosis of HRM effectiveness and human centeredness by aligning strategic objectives
with employee experiences.

For example, in recruiting, digital platforms are key to reducing the time to hire and expand
access to a more diverse talent pool, but these operational gains must be assessed in tandem
with candidate perceptions of fairness and subsequent alignment between selected
employees and organizational culture. Similarly, blended training programs that combine
online modules with in-person workshops can increase flexibility and scalability, but their
Pag. 37 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

real impact is reflected in employee satisfaction, skill acquisition, and the applicability of
learning to everyday work contexts.

Performance management further exemplifies this dualism. While digital performance

dashboards provide valuable data on goal tracking and productivity, they must be
complemented by human-centered feedback processes to ensure clarity, transparency, and
motivational support. Digital communication platforms, such as Slack or Microsoft Teams,
facilitate real-time interaction and documentation, but without sustained interpersonal
dialog, such as one-on-one meetings or structured team discussions, the risk of alienation
or miscommunication increases (Table 3).

HRM Organizatio Employee Measureme Measureme Result

Strategy nal KPIs KPIs nt Method nt method indicator
(Org) (employee) and
thresholds

Recruitme Time to hire, Perceived ATS Onboarding Employme

nt cost- fairness, fit, metrics, surveys, nt time <
effectiveness, and recruitment feedback 30 days,
diversity of commitment cost reports for new diversity
talent after employees ratio >
employment 30%

Training Return on Satisfaction, Training Post- ROI >

and investment in applicability, ROI training 150%,
developme learning, perception of Calculator, assessment, completion
nt completion growth LMS competence rate > 80%
rates, Analysis self-
improvement assessment
indices

Performan Goal Transparency Dashboard 360° Achievem

ce alignment , clarity, reports, KPI feedback, ent > 85%,
Manageme indicators, motivation completion pulse positive
nt feedback and rates surveys feedback >
cycle assumption 75%
efficiency of objectives

Labour Speed of Trust in HR, HR system Engagemen Problem

relations resolution, quality of logs, t survey, solving < 3
policy communicati HR days,

Pag. 38 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

compliance, on, emotional compliance satisfaction engageme

digital climate audits rating nt score >
adoption 70%

Digital Adoption Digital trust, Adoption Digital Adoption

integration rate, digital autonomy, dashboards, skills self- rate >
productivity, learning usage logs checks, user 90%,
problem curve experience satisfactio
solving surveys n with
tools >
80%

Process Reduction of Accessibility, Process Employee Reduction

automatio administratio customizatio logs, HR feedback of working
n n burdens, n, technical operations forms, time >
time stress level reports usage 30%, error
efficiency, ratings rate < 2%
error rate

Predictive Forecast Perception of Predictive Perception, Prediction

analytics accuracy, fairness, trust accuracy Trust Index accuracy >
improved in data, reports, and 85%, wear
retention, alignment retention Fairness < 10%
talent with statistics Surveys
utilization aspirations

Table 3 Detailed HRM Assessment Matrix

To assess the effectiveness of these strategies in a structured and multidimensional manner,

we propose an HRM evaluation matrix that captures both the macro- and micro-level
dimensions. On the organizational side, the matrix tracks metrics, such as recruitment cycle
time, training ROI, alignment between individual and corporate goals, digital adoption
rates, administrative workload reduction, and predictive accuracy. In terms of employees,
the model captures perceptions of fairness, engagement, satisfaction, digital competency,
learning experience, and overall alignment with personal and professional goals.

This framework provides a strategic tool for HR professionals and decision makers to assess
the real-world impact of hybrid HR management strategies. It promotes a balanced
perspective, emphasizing that technological progress should not come at the expense of
human connections, nor should traditional methods resist innovation for the sake of
familiarity. Instead, an adaptive HRM system should continually assess and adjust its

Pag. 39 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

components based on data-driven insights and human feedback to ensure resilience,

relevance, and sustainability.

In addition, the adoption of digital tools should precede and be supported by targeted
training and capacity-building efforts to address potential resistance and digital skills gaps.
Organizations, such as Microsoft, have demonstrated the effectiveness of a leadership-first
model, in which managers are the first to be trained in the use of new technologies, ensuring
smoother transitions and internal advocacy. In parallel, the automation of repetitive HR
tasks, such as payroll, scheduling, and documentation, can significantly improve
operational efficiency, but only when counterbalanced by the continued presence of HR
professionals who provide developmental guidance and psychological support.

Advanced HR analytics, such as those deployed by IBM, allows the prediction of employee
disengagement or turnover risk, enabling timely and personalized interventions. However,
the use of such tools needs to be governed by ethical principles and transparent
communication to foster trust among employees and to avoid perceptions of surveillance
or misuse of data.

Figure 2 Human resource management assessment framework at organization and

employee level

Pag. 40 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 3 Example of Evaluation of Two-Matrix Evaluation

6. Conclusions

A comparative analysis between traditional and digitally enhanced human resource

management (HRM) methods reveals fundamental distinctions in the way organizations
approach employee-related processes and operational efficiency. While both models have
demonstrated value in different organizational contexts, this study highlights the increasing
relevance of integrating digital technologies into HRM frameworks to meet growing
demands for scale, speed, personalization, and data security.

Traditional HRM methods tend to be slower, more rigid, and less scalable, although they
are based on interpersonal interactions and established routines. Their reliance on manual
operations makes them less suitable for large- or medium-sized organizations that require
real-time feedback. However, they offer a sense of familiarity and trust, especially in
smaller institutions or cultures, where face-to-face communication remains the norm.

By contrast, digital HRM methods offer significant advantages in terms of automation,

speed, and adaptability. Tools such as artificial-intelligence-based recruitment software, e-
learning platforms, performance analytics, and cloud-based HR systems enable
organizations to streamline processes while delivering more personalized and accessible
employee experiences. This increased flexibility contributes to greater satisfaction and
engagement, particularly in hybrid or remote working environments.
Pag. 41 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

However, digital transformation in HRM is challenging. The initial implementation requires

considerable investment in infrastructure, training, and change management. Security and
data protection, especially under frameworks such as GDPR, must be proactively
addressed. Despite these challenges, properly managed digital systems can provide greater
control over sensitive data and improve decision-making accuracy.

A key insight of this study is the complementary nature of these two approaches. Rather
than viewing traditional and digital HRM as mutually exclusive, organizations should seek
to combine their strengths. For example, while digital tools can optimize recruitment and
training, performance reviews can benefit from face-to-face interactions to maintain trust
and ensure nuanced communication.

This study also highlights the growing importance of predictive analytics and machine
learning in HRM. Organizations, such as Accenture and Microsoft, have already begun to
implement models capable of identifying employees at risk of disengagement or having
difficulty adopting new technologies. This information enables targeted interventions such
as customized training or mentoring programs, thereby promoting retention and
adaptability. Predictive models trained on historical data and employee feedback can predict
the likelihood of attrition or development gaps, enabling a more strategic and proactive
approach to HR.

From a methodological perspective, the proposed two-level assessment framework,

focusing simultaneously on organizational performance indicators and employee
perceptions, represents a significant contribution to the field. This two-dimensional model
enables a nuanced assessment of HRM initiatives, supporting the alignment of strategic
outcomes with people-centered values.

Despite its practical relevance, this study has some limitations. First, the assessment is
based on secondary data and theoretical models, which may not capture the full complexity
of HR dynamics across industries and cultures. Second, the predictive potential of digital
tools is limited by the quality and granularity of available data. Third, although examples
from companies such as Google, IBM, and Spotify provide valuable benchmarks,
generalization to smaller organizations may require additional contextual adaptation.

Future research should prioritize the empirical validation of the proposed framework
through longitudinal case studies and sector-specific analyses. In addition, it is important
to carefully consider the ethical concerns regarding the use of AI in HRM, particularly
regarding transparency, openness, and employee autonomy.

Based on the findings and framework presented, several practical recommendations have
emerged. First, organizations are encouraged to pursue the strategic integration of
traditional and digital HR methods, tailoring their application to different stages of the
employee lifecycle. Second, it is advisable to gradually implement digital tools, allowing
Pag. 42 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ample time for organizational adaptation. Finally, sustained investments in employee

training and enhanced digital literacy are vital for fully leveraging HR technology
investments and mitigating resistance to change.

In short, hybridizing HRM practices - if guided by robust assessment, ethical technology

implementation, and sustained attention to employee experience - can enable organizations
to navigate the complexities of the digital age while reinforcing human value at the core of
organizational development.

References
[1] Agarwal, R., Gautham, L., & Choudhury, P. (2020). The role of human capital in the
digital economy: an empirical investigation. Journal of Management, 46(3), 561-579.

[2] Bass, B. M., & Riggio, R. E. (2021). Transformational leadership. Psychology press.

[3]Deci, E. L., & Ryan, R. M. (2000a). Self-Determination Theory and the Facilitation of
Intrinsic Motivation, Social Development, and Well-Being.

[4] DeNisi, A. S., & Williams, K. J.(2020). The nature of feedback in the
workplace. Oxford Handbook of Organizational Psychology. Oxford University
Press.

[5] Gartner. (2023). Gartner's HR Technology Market Guide.

[6] Greenhaus, J. H., & Powell, G. N. (2020). Work and family: theory and research.
Psychology press.

[7] Guthrie, J. P., & Olian, J. D. (2021). Strategic alignment of HRM practices and
organizational performance: emerging trends. Journal of Strategic HRM, 15(2), 137-154.

[8] IBM. (2023). AI in Human Resources: Transforming Recruitment and

Performance Management.

[9] Jarrahi, M. H. (2020). The adoption of digital HR technologies in organizations: a

review and future directions. Journal of Business Research, 112, 11-23.

[10] Lawler, E. E., & Boudreau, J. W. (2021). Effective Human Resource

Management: A Global Analysis. Stanford University Press.

Pag. 43 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[11] LinkedIn. (2023). Global Recruitment Trends 2023.

[12] Noah, R. A. (2020). Employee training and development (7th edition).

McGraw-Hill Education.

[13] Ployhart, R. E., & Bliese, P. D. (2020). Building and developing human capital in
organizations: the role of adaptability and innovation. Annals of the Academy of
Management, 14(2), 1-29.

[14] Rahim, M.A. (2002) Toward a Theory of Managing Organizational Conflict.

[15] Shore, L. M., Cleveland, J. N., & Sanchez, D. (2021). Inclusive Workplaces: The
Role of HRM in Building Diversity and Equity. Human Resource Management Review,
31(3), 100-115.

Bibliography
[1] Ahmad, S., & Schroeder, R. G. (2007). "The value of human resource management for
organizational performance" – Business Horizons, 50(6), 503-511.

[2] Green, K. W., Wu, C., Whitten, D., & Medlin, B. (2011). "The impact of strategic
human resource management on firm performance and HR professionals' work attitude
and work performance" – The International Journal of Human Resource Management,
17(4), 559-579.

[3] Kareem, M. A. (2015). "The Impact of Human Resource Management Strategies on

Organizational Performance" – International Journal of Management and Commerce
Innovations, 3(2), 1-7.

[4] Liao, Y. (2005). "Business strategy and performance: The role of human resource
management control" – Personnel Review, 34(3), 294-309.

[5] Perez Lopez, S., Montes Peon, J. M., & Vazquez Ordas, C. J. (2005). "Human
Resource Practices, Organizational Learning and Business Performance" – Human
Resource Development International, 8(2), 147-164.

Pag. 44 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

MEASURING THE PERFORMANCE OF A SMART HOME AUTOMATION

SOFTWARE USING DESIGN PATTERNS

Tudor-Călin CIOT1
Ştefan BUTACU2
Costin-Anton BOIANGIU3
Cătălin TUDOSE4

Abstract
In the rapidly evolving landscape of smart home technologies, the efficiency of software
systems is paramount in ensuring seamless automation experiences for users. This scientific
article proposes an in-depth examination of the performance metrics of a Smart Home
Automation System (SHAS) software constructed through the integration of diverse design
patterns. The study aims to contribute valuable insights into the impact of design patterns
on the efficiency, reliability, and scalability of the SHAS software.
The research employs a rigorous methodology that encompasses the identification,
implementation, and analysis of various design patterns within the software architecture.
Key performance indicators such as response time, resource utilization, and system
scalability will be systematically evaluated to assess the overall effectiveness of the chosen
design patterns. Comparative analyses will be conducted to highlight the advantages and
potential challenges associated with each pattern.
Through this investigation, we anticipate uncovering optimal design patterns that enhance
the SHAS software's performance, ultimately contributing to the advancement of smart
home technologies. The findings of this study hold significant implications for developers,
researchers, and industries engaged in the design and implementation of intelligent home
automation systems.
Keywords: software development, design patterns, Internet of Things
JEL Classification: C61

1Student, POLITEHNICA National University for Science and Technology of Bucharest, Romania,
tudorcalin.ciot@stud.acs.pub.ro
2Student, POLITEHNICA National University for Science and Technology of Bucharest, Romania,

stefan.butacu@stud.acs.pub.ro
3PhD, Professor, POLITEHNICA National University for Science and Technology of Bucharest, Romania,

costin.boiangiu@cs.pub.ro, corresponding author

4PhD, Lecturer, POLITEHNICA National University for Science and Technology of Bucharest and Luxoft

Romania, Romania, catalin.tudose@gmail.com

Pag. 45 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1. Introduction
In the contemporary era of smart living, Smart Home Automation Systems (SHAS) have
emerged as integral components, reshaping the way we interact with and manage our living
spaces. These systems leverage cutting-edge technologies to provide users with
unprecedented control over various aspects of their homes, from lighting and climate to
security and entertainment. At the heart of these systems lies the software infrastructure, a
critical determinant of the overall performance and user experience.
As the demand for smart home solutions burgeons, the imperative to develop efficient and
scalable SHAS software becomes increasingly paramount. The efficacy of such software is
intricately tied to the underlying architectural decisions, with design patterns playing a
pivotal role in shaping the software's structure, modularity, and extensibility. This scientific
article embarks on a comprehensive exploration into the performance metrics of a Smart
Home Automation System software, specifically designed and implemented with a variety
of design patterns.
The motivation for this study stems from the recognition that while design patterns offer
proven solutions to recurring architectural challenges, their impact on the performance of
SHAS software remains a relatively underexplored domain. By dissecting and evaluating
the performance implications of various design patterns, this research aims to provide a
nuanced understanding of their role in enhancing or potentially impeding the overall
functionality of smart home automation.
Through meticulous analysis and empirical measurements, we seek to shed light on the
intricate relationship between design patterns and the performance attributes crucial to
SHAS software, including responsiveness, resource utilization, and scalability. This
investigation is poised to unravel insights that not only contribute to the academic discourse
on software architecture but also offer practical guidance to developers and industry
stakeholders engaged in the evolution of intelligent home automation systems.
In essence, this study endeavors to bridge the gap between theoretical design paradigms
and real-world performance outcomes, fostering a deeper comprehension of the intricate
interplay between design patterns and the efficacy of Smart Home Automation System
software.
The landscape of smart home technologies is characterized by an ever-expanding array of
devices, protocols, and user preferences. Consequently, the need for SHAS software to
seamlessly adapt to this complexity underscores the significance of selecting appropriate
design patterns. However, while design patterns are recognized for their ability to enhance
software maintainability and flexibility, their influence on performance remains a dynamic
field of investigation [1].
This study acknowledges the dynamic nature of smart home environments and the necessity
for SHAS software to not only accommodate diverse functionalities but also to execute
Pag. 46 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

these operations with optimal efficiency. As the smart home ecosystem evolves, the role of
SHAS software becomes increasingly intricate, demanding a meticulous evaluation of the
impact of design patterns on its performance characteristics.
Through a structured examination of the chosen design patterns, our research seeks to
address fundamental questions surrounding their efficacy in the context of SHAS software.
Which design patterns prove most effective in optimizing response times? How do different
patterns impact resource utilization, and to what extent do they contribute to or alleviate
scalability challenges inherent in smart home environments? These inquiries form the crux
of our investigation and aim to elucidate the nuanced relationships between design
decisions and the tangible performance outcomes in SHAS software.
In a rapidly advancing technological landscape, the findings of this research are poised to
inform not only the development of smart home automation systems but also the broader
discourse on the symbiotic relationship between software architecture and performance
optimization. By navigating the intricate terrain where design patterns intersect with the
demands of modern smart living, this study aspires to furnish valuable insights for
architects, developers, and researchers committed to advancing the frontiers of smart home
technologies.
The concept of smart home automation systems has evolved over several decades, blending
technological advancements with the vision of creating more convenient and efficient living
spaces. The history of smart home automation can be traced back to early attempts at
automating household tasks and integrating technology into homes [2].
The roots of smart home automation can be found in the 1950s and 1960s, marked by the
introduction of basic home automation concepts. At this time, futuristic visions of homes
equipped with automated appliances and systems began to emerge. The idea was often
portrayed in science fiction literature and films, influencing the public's perception of what
the future home might look like.
The 1980s and 1990s witnessed the development of various home automation protocols and
systems. X10, one of the earliest home automation protocols, allowed devices to
communicate over power lines. This technology laid the foundation for controlling lights,
appliances, and other devices remotely.
Despite the innovations, adoption remained limited due to interoperability issues and a lack
of standardized communication protocols. Home automation systems during this period
were often expensive, complex, and accessible only to enthusiasts or those with substantial
financial resources.
The 2000s marked a significant turning point for smart home automation, driven by
advancements in connectivity and the proliferation of the internet. The rise of Wi-Fi
technology and the development of the Internet of Things (IoT) paved the way for a more
interconnected and accessible smart home ecosystem.
Pag. 47 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Smart home devices and systems began to gain popularity, offering enhanced control and
monitoring capabilities. Companies introduced products like smart thermostats, security
cameras, and automated lighting systems. The advent of smartphones played a pivotal role,
providing users with the ability to control their homes remotely through dedicated apps.
The 2010s witnessed a surge in the adoption of smart home technologies, moving beyond
individual devices to integrated ecosystems. Major tech companies introduced
comprehensive platforms, such as Apple's HomeKit, Google's Nest, and Amazon's Alexa,
aiming to streamline the user experience and enhance interoperability.
Voice-activated assistants became a common feature in smart homes, allowing users to
control devices through natural language commands. Integration with third-party services
and devices further expanded the capabilities of smart home automation systems.
In the present day, smart home automation has become increasingly ubiquitous, with a wide
range of devices and systems available to consumers. The integration of artificial
intelligence (AI) and machine learning has enhanced the intelligence and adaptability of
smart home systems, enabling them to learn user preferences and anticipate needs.
The future of smart home automation holds the promise of even greater connectivity,
interoperability, and energy efficiency. As technology continues to advance, smart homes
are likely to evolve into highly adaptive environments that seamlessly integrate with other
aspects of daily life, contributing to a more sustainable and comfortable living experience
and maybe even integrating robots in the process.
Robots in homes and their integration with smart home automation systems hold great
promise for transforming our living spaces. As technology continues to advance, robots are
poised to play increasingly integral roles in enhancing convenience, security, and overall
home management. Here are some key aspects of the future trajectory for robots in homes
and their integration with smart home automation systems: personal assistance and
companionship, smart home integration, autonomous cleaning and maintenance, security
and surveillance, healthcare assistance, customized environmental control, education and
entertainment, and energy efficiency.
While the future of robots in homes and their integration with smart home automation
systems presents exciting possibilities, ethical considerations, and privacy concerns will
also need to be addressed. Striking a balance between innovation and responsible
deployment will be crucial for ensuring the widespread acceptance and success of these
technologies in our homes.

Pag. 48 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

2. Related Work
Smart home automation systems rely on a variety of hardware platforms to connect and
control various devices. These platforms can be broadly categorized into dedicated smart
home hubs, smartphone/tablet apps, and web-based interfaces.
Devoted smart home hubs serve as the central control unit, managing communication with
connected devices. They often feature voice assistants, local processing capabilities, and
customizable automation rules. Popular examples include Amazon Echo, Google Home,
and Samsung SmartThings.
Smartphone and tablet apps provide a convenient way to manage smart home systems
remotely. They offer a user-friendly interface for controlling devices, creating automation
rules, and accessing real-time sensor data. Examples include the official apps for various
smart home hubs and standalone apps for specific devices or functions.
Web-based interfaces offer a platform-independent way to manage smart home systems.
They provide access to all the features and capabilities of the system from any web browser.
This can be particularly useful for managing devices from a computer or when using
multiple devices [3].
The related work in the field of smart home automation systems with a focus on employing
design patterns for performance improvement encompasses a range of seminal books,
articles, and research papers.
The foundational work of [4] establishes a solid understanding of design patterns that can
be applied to enhance the modularity and maintainability of smart home automation
software. Additionally, "Building Scalable and High-Performance Java Web Applications
Using J2EE Technology" by Greg Barish contributes valuable insights into scalable
software design, offering principles applicable to the performance optimization of smart
home systems.
Extensions of these principles to enterprise-level applications provide a framework for
designing scalable and robust systems that align with the complex requirements often found
in smart home environments [5]. Furthermore, [6] addresses the broader context of IoT,
providing a foundation for understanding the architectural considerations and protocols
relevant to smart home automation.
While technical, [7] provides practical guidance on applying UML and design patterns in
software development. The insights are particularly valuable for iterative development
processes, aligning with the dynamic and evolving nature of smart home automation
systems.
In parallel, works such as [8] offer a user-centric perspective, emphasizing the importance
of understanding end-user needs and experiences in the design and implementation of smart
home automation software. This user-focused approach complements the technical

Pag. 49 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

literature, providing a well-rounded understanding of the challenges and opportunities in

the domain.
To stay current with the latest advancements, researchers often turn to the IEEE Xplore
Digital Library and the ACM Digital Library, which host a plethora of research papers.
Exploring these databases with keywords such as "smart home," "IoT," "design patterns,"
and "performance" yields a wealth of recent research, offering novel approaches and
techniques for optimizing smart home automation software. The collective body of related
work provides a comprehensive foundation for the design and evaluation of smart home
automation systems leveraging design patterns for enhanced performance.
Continuing in the realm of related work, recent advancements in smart home automation
systems and design patterns have been shaped by a dynamic landscape of research and
development. The exploration of cutting-edge concepts often involves perusing the latest
publications available in scholarly databases and forums.
One notable source of inspiration is the IEEE Xplore Digital Library, where researchers
delve into a multitude of papers to stay abreast of evolving methodologies. The integration
of design patterns into the development of smart home automation software is often
informed by the findings of studies such as [9]. This survey not only provides a historical
perspective but also sheds light on the diverse applications of IoT, a critical context for
understanding the interconnected nature of smart home ecosystems.
Furthermore, recent research articles [10] explore the integration of edge computing to
improve the real-time processing capabilities of smart home automation software. This
represents a novel extension to traditional design patterns, considering the distributed nature
of computation in modern smart home environments.
In the pursuit of optimizing energy efficiency, [11] introduces green computing principles
to the realm of smart homes. This work is particularly relevant for those seeking to design
sustainable and eco-friendly smart home automation solutions.
Additionally, researchers interested in the security aspects of smart home automation
software can refer to [12]. This survey not only outlines the existing security challenges but
also provides insights into incorporating secure design patterns to fortify smart home
systems against potential threats.
Collaborative efforts, as demonstrated in [13], offer insights into cooperative security
models that leverage design patterns to enhance the resilience of smart home automation
systems. The cooperative approach acknowledges the interconnected nature of devices
within a smart home and proposes strategies for collaborative threat detection and
mitigation.
In conclusion, the related work in the domain of smart home automation software,
especially concerning the integration of design patterns for performance enhancement,
encompasses a rich and evolving body of literature. The combination of foundational
Pag. 50 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

principles from classic texts, insights from recent publications, and emerging paradigms in
areas such as edge computing, green computing, and cooperative security contribute to a
holistic understanding of the challenges and opportunities in this dynamic field.

3. Proposed Architecture
The proliferation of smart home technologies has ushered in an era where homes are
increasingly equipped with a myriad of interconnected devices, ranging from thermostats
and lighting systems to security cameras and entertainment units. As the complexity and
diversity of smart home environments continue to expand, the demand for intelligent
automation systems capable of orchestrating these devices seamlessly has become
paramount. To address the challenges inherent in managing such intricate ecosystems, this
scientific article introduces a sophisticated architectural framework that leverages well-
established design patterns.
Smart home automation systems play a pivotal role in enhancing user comfort, security, and
energy efficiency. However, as the number and diversity of devices within these systems
grow, so does the complexity of managing and coordinating their interactions. This
complexity necessitates innovative approaches to software architecture that can
accommodate the dynamic nature of smart home environments while providing a robust
foundation for scalability, adaptability, and ease of maintenance.
In response to these challenges, our proposed architectural framework embraces a carefully
curated set of design patterns. These design patterns serve as building blocks, each
addressing specific concerns critical to the success of a smart home automation system. By
integrating the Singleton Pattern, Observer Pattern, Command Pattern, Factory Pattern, and
Decorator Pattern, our framework aims to provide a comprehensive solution that addresses
fundamental aspects such as configuration management, real-time device monitoring,
customizable automation tasks, dynamic device creation, and extensible feature
augmentation.
The use of design patterns in software architecture is a well-established practice, and their
relevance becomes particularly pronounced in the context of smart home automation. As
homes evolve into intelligent ecosystems, the need for a centralized configuration manager
(utilizing the Singleton Pattern) becomes evident to ensure consistency across diverse
devices. Real-time updates on device states, facilitated by the Observer Pattern, are essential
for users to stay informed and maintain control over their smart home environment.
Furthermore, the Command Pattern empowers users with the ability to create and execute
commands, providing a user-friendly interface for customization. The Factory Pattern
addresses the challenge of accommodating various device types by introducing modularity
and adaptability to the system. Finally, the Decorator Pattern facilitates the dynamic
augmentation of device functionalities, enabling the smart home automation system to
evolve alongside technological advancements and changing user needs.
Pag. 51 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1. Singleton Pattern: Centralized Configuration Manager

The Singleton Pattern [4] plays a pivotal role in ensuring the consistency and coherence of
configuration settings throughout the smart home automation system. By implementing a
centralized configuration manager as a singleton, we guarantee that there exists only one
instance responsible for managing configuration parameters. This design choice facilitates
a unified point of access for configuration settings across the entire application. Whether
it's regulating the behavior of individual devices or establishing system-wide preferences,
the Singleton Pattern ensures a single, authoritative source for configuration data. This
approach simplifies maintenance, reduces the likelihood of conflicting configurations, and
enhances the overall reliability of the smart home automation system.

2. Observer Pattern: Real-Time Device Monitoring

In the context of a smart home automation system, real-time updates on the state of devices
are crucial for providing users with accurate information and facilitating prompt decision-
making. The Observer Pattern [4] is employed to establish a dynamic communication
mechanism between smart devices and the user interface. Each smart device serves as a
subject, and the user interface acts as the observer. When the state of a device changes, it
notifies the observer (user interface) instantly, allowing for real-time updates on the user
interface. This ensures that users are well-informed about the status of their smart home
devices and can take immediate actions based on the current conditions.

3. Command Pattern: Customizable Automation Tasks

The Command Pattern [4] empowers users with a flexible and intuitive means of
customizing automation tasks within the smart home environment. Users can create
commands encapsulating specific operations and execute them as needed. This pattern
facilitates the decoupling of the sender (user interface or automation controller) from the
receiver (smart device), enabling a wide range of customization possibilities. Whether it's
scheduling routines, automating sequences of actions, or responding to specific events, the
Command Pattern provides a versatile framework for users to tailor the smart home
automation system to their preferences.

4. Factory Pattern: Dynamic Device Creation

The Factory Pattern [4] is instrumental in addressing the diverse landscape of smart devices
within a modern home. By creating various smart device factories, we establish a modular
approach and each factory is responsible for creating a specific type of device, ensuring that
the system can seamlessly incorporate new device types without modifying existing code.
Pag. 52 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

This not only simplifies the addition of new devices but also enhances the system's
adaptability to emerging technologies. The Factory Pattern contributes to the scalability and
maintainability of the smart home automation system by promoting a consistent and
structured approach to device creation.
Each factory is responsible for creating a specific type of device, ensuring that the system
can seamlessly incorporate new device types without modifying existing code. This not
only simplifies the addition of new devices but also enhances the system's adaptability to
emerging technologies. The Factory Pattern contributes to the scalability and
maintainability of the smart home automation system by promoting a consistent and
structured approach to device creation.

4. Deployment Tools
The burgeoning field of smart home automation demands meticulous consideration of
deployment tools to orchestrate the seamless integration and optimal functionality of
diverse hardware and software components. This part of the research endeavors to
scrutinize and assess the preeminent deployment tools prevalent in the smart home
automation domain, examining key attributes such as compatibility, user-friendliness,
community support, and customization capabilities. As smart homes become increasingly
sophisticated, the choice of an appropriate deployment tool plays a pivotal role in ensuring
the reliability and efficiency of these systems.
Among the many deployment tools, Home Assistant emerges as a prominent open-source
platform renowned for its versatility and expansive device compatibility. With deployment
options ranging from the resource-efficient Raspberry Pi to the scalable Docker containers,
Home Assistant caters to users with diverse hardware preferences. Its commitment to local
control and privacy underscores its appeal, offering users a robust foundation for
orchestrating a comprehensive smart home ecosystem.
In the realm of Java-based deployment tools, Open Home Automation Bus (OpenHAB)
distinguishes itself by prioritizing protocol compatibility and cross-platform versatility.
OpenHAB's capacity to seamlessly operate across Windows, Linux, and macOS positions
it as an attractive choice for users seeking a comprehensive solution that accommodates a
spectrum of smart home technologies. Its robust framework and active community
contribute to its standing as a formidable deployment option.
Built on Node.js, ioBroker offers an open-source automation platform designed to
harmonize with diverse hardware environments. Its support for various devices, coupled
with cross-platform compatibility, provides users with a flexible and adaptable deployment
solution. IoBroker's modular architecture empowers users to integrate and control an array
of smart home devices, establishing it as a dynamic player in the smart home automation
landscape.
Pag. 53 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Recognized for its lightweight design and ease of use, Domoticz positions itself as a
pragmatic choice for users entering the realm of home automation. Supporting an array of
devices and protocols, Domoticz offers deployment options across platforms like Raspberry
Pi, Windows, and Linux. Its accessibility and broad compatibility contribute to its appeal
as a straightforward yet powerful deployment tool.
For those inclined towards visual programming, Node-RED [14] stands out as a compelling
tool tailored for IoT and home automation projects. Its node-based interface simplifies the
creation of automation flows, allowing users to design intricate smart home scenarios.
Node-RED's deployment across various platforms enhances its accessibility and integration
capabilities.
In the paradigm of containerization, Docker [15] emerges as a cornerstone for efficient and
portable smart home automation deployment. Docker containers provide a lightweight and
standardized environment, facilitating the deployment of software across diverse platforms.
Platforms such as Home Assistant and OpenHAB offer official Docker images,
underscoring Docker's role in streamlining the deployment process.
This comparative analysis of deployment tools within the smart home automation domain
illuminates the diverse landscape of options available to users and developers. Each tool
presents a unique set of features and advantages, catering to different preferences and
requirements. By considering factors such as device compatibility, deployment flexibility,
and community support, stakeholders can make informed decisions when selecting the most
suitable deployment tool for their smart home automation endeavors. As the field continues
to evolve, ongoing research will be paramount in assessing emerging tools and
technologies, ensuring that smart home deployments remain at the forefront of innovation
and efficiency.

5. Elements of Comparison
As the landscape of software development continues to evolve, the importance of
maintaining code quality remains paramount. Object-oriented programming, a widely
adopted paradigm, emphasizes modular design and code reusability. In this context, metrics
play a crucial role in assessing various aspects of code quality, as we described in the
analysis from Table 1.
Table 1

Free Memory (in KB)

With Design Patterns Without Design Patterns

248,952 143,154

Pag. 54 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

In the pursuit of advancing our understanding of these metrics and their implications on
code quality, this paper presents the outcomes of rigorous experiments conducted on SHAS
software. The experiments aim to validate the efficiency of these metrics in assessing the
quality of object-oriented code. For example, we obtained a better memory performance
when using the design patterns as we see in Table 2. Moreover, the study investigates the
relationships between these metrics and their combined impact on the overall
maintainability and performance of a codebase.

Table 2

Static Code Analysis (Average Values)

Classes CBO LOC LCOM* TCC
25 3 62 05875 0.7825

CBO (Coupling between objects): Counts the number of dependencies a class has. The
tools check for any type used in the entire class (field declaration, method return types,
variable declarations, etc). It ignores dependencies to Java itself (e.g. java.lang.String).
LOC (Lines of code): It counts the lines of code, ignoring empty lines and comments. The
number of lines here might be a bit different from the original file, as we use Eclipse JDT
(Java Development Tools) internal representation of the source code to calculate it.
LCOM* (Lack of Cohesion of Methods): This metric is a modified version of the current
version of LCOM implemented in CK Tool. LCOM* is a normalized metric that computes
the lack of cohesion of class within a range of 0 to 1. Then, the closer to 1 the value of
LCOM* in a class, the less the cohesion degree of this respective class. The closer the value
of LCOM* in a class, the greater the cohesion of this respective class. This implementation
follows the third version of LCOM* defined in [16].
TCC (Tight Class Cohesion): Measures the cohesion of a class with a value range from 0
to 1. TCC measures the cohesion of a class via direct connections between visible methods,
two methods or their invocation trees access the same class variable.

6. Conclusions and Future Work

In conclusion, this research paper has undertaken a comprehensive investigation into the
performance of a Smart Home Automation System Software implemented in Java and
designed with the incorporation of various object-oriented design patterns. The findings and
analyses presented herein shed light on key aspects that contribute to the effectiveness and
efficiency of the software in real-world smart home environments.
Pag. 55 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Through rigorous performance testing and measurement, we have evaluated critical aspects
such as response time, throughput, scalability, resource utilization, and concurrency. These
metrics provide a holistic view of how well the system responds to user interactions and
handles varying workloads.
The integration of design patterns, such as Singleton, Observer, and Command, has played
a pivotal role in enhancing the software's architecture. The application of these patterns has
improved modularity, flexibility, and maintainability, contributing to the overall success of
the system.
The adherence to object-oriented principles in Java, including encapsulation, inheritance,
and polymorphism, has been instrumental in achieving a well-structured and extensible
codebase. This has facilitated easier maintenance and future enhancements to the Smart
Home Automation System Software.
The research has highlighted the importance of design patterns in managing device
compatibility, ensuring that the software seamlessly integrates with diverse smart devices.
This interoperability is essential for creating a cohesive and user-friendly smart home
ecosystem.
Building on the insights gained from this research, several avenues for future work and
improvement have been identified.
Investigate the incorporation of additional design patterns to further optimize specific
aspects of the software, addressing any identified performance bottlenecks or enhancing
specific functionalities.
Explore the integration of machine learning algorithms to predict user behavior and
automate device control based on historical usage patterns. This could contribute to a more
intelligent and adaptive smart home system.
Investigate the potential benefits of a distributed architecture, utilizing design patterns
suitable for distributed systems. This could enhance scalability and resilience in large-scale
smart home deployments.
Implement design patterns to facilitate dynamic device discovery and configuration,
allowing the system to seamlessly adapt to changes in the smart home environment without
manual intervention.
Research and apply design patterns that promote energy efficiency, considering the growing
emphasis on sustainable and eco-friendly smart home solutions.
Conduct real-world deployment studies to validate the software's performance in diverse
smart home environments. This includes considering factors such as network conditions,
user behaviors, and the presence of various smart devices.

Pag. 56 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Foster community involvement and collaboration by open-sourcing certain components or

providing APIs that encourage third-party developers to contribute to the software's
evolution.
Implement continuous performance monitoring mechanisms to detect and address any
performance degradation over time. This ensures that the software remains optimized and
responsive as it evolves.
By addressing these areas in future work, the Smart Home Automation System Software
can evolve into a more sophisticated and adaptive solution, continually meeting the ever-
changing demands of the smart home ecosystem. This research lays the foundation for
ongoing advancements in the field, contributing to the broader discourse on the intersection
of Java, design patterns, and performance in smart home automation systems.

References
[1] D. Community, "Domoticz: Open-source home automation system".
[2] O. Community, "Open Home Automation Bus (OpenHAB)" 2021.
[3] A. RAJAGOPALAN, C. SUNDARARAJAN Smart Home Automation Systems: A
Survey of Hardware Platforms and Software Architectures, 2019.
[4] E. GAMMA, R. HELM, R. JOHNSON, J. VLISSIDES Design Patterns: Elements of
Reusable Object-Oriented Software, 1994.
[5] M. FOWLER Patterns of Enterprise Application Architecture, 2002.
[6] S. CIRANI, G. FERRARI, M. PICONE, L. VELTRI Internet of Things (IoT)
Architectures, Protocols, and Standards, 2018.
[7] C. LARMAN Applying UML and Patterns: An Introduction to Object-Oriented Analysis
and Design and Iterative Development, 2004.
[8] D. SPIVEY Home Automation For Dummies, 2015.
[9] S. TANWAR, S. TYAGI A Comprehensive Survey on Internet of Things (IoT) from 2008
till 2019, 2019.
[10] M. YANG, L. ZHOU Enhancing the Performance of Smart Home Systems through
Edge Computing, 2020.
[11] H. T. EL KASSABI, M. M. MORSY Green IoT: An Investigation on the Role of Edge
Computing, 2020.
[12] A. G. AMEEN, S. AL-KINDI Security and Privacy Issues in IoT-Based Smart Home:
A Comprehensive Survey, 2020.

Pag. 57 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[13] S. ANWAR Towards Cooperative Security for IoT-Based Smart Home: A Survey, 2021.
[14] Node-RED, A visual tool for wiring the Internet of Things, https://nodered.org/.
[15] Docker - Build, Share, and Run Any App, Anywhere, https://www.docker.com/.
[16] B. HENDERSON-SELLERS, L. CONSTANTINE, I. GRAHAM Coupling and
cohesion (towards a valid metrics suite for object-oriented analysis and design) vol. Object
Oriented Systems 3, 1996.

Bibliography
S. GREENGARD The Internet of Things, revised and updated edition (The MIT Press
Essential Knowledge series), 2021
F.J. DIAN Fundamentals of Internet of Things: For Students and Professionals, 2022
M. SONI, J. SONI Internet of Things Textbook: Transforming Everyday Life with
Technology (Internet of Things books), 2024
E. GAMMA, R. HELM, R. JOHNSON, J. VLISSIDES Design Patterns: Elements of
Reusable Object-Oriented Software, 1994
E. FREEMAN, E. ROBSON, B. BATES, K. SIERRA Head First Design Patterns: A Brain
Friendly Guide 2nd Edition, 2004
E. GAMMA, R. HELM, R. JOHNSON, J. VLISSIDES Design Patterns: Elements of
Reusable Object-Oriented Software, 2015
E. FREEMAN, E. ROBSON Head First Design Patterns: Building Extensible and
Maintainable Object-Oriented Software 2nd Edition, 2021
R.C. MARTIN Clean Architecture: A Craftsman's Guide to Software Structure and Design
(Robert C. Martin Series), 2017
A. SHALLOWAY, J. TROTT Design Patterns Explained: A New Perspective on Object
Oriented Design, 2nd Edition (Software Patterns), 2004
J.J. GEEWAX API Design Patterns, Manning, 2021
B. RUECKER Practical Process Automation: Orchestration and Integration in
Microservices and Cloud Native Architectures, O'Reilly Media, 2021
A. AXELROD Complete Guide to Test Automation: Techniques, Practices, and Patterns
for Building and Maintaining Effective Software Projects, Apress, 2018
A. SWEIGART Automate the Boring Stuff with Python, 2nd Edition: Practical
Programming for Total Beginners, No Starch Press, 2019

Pag. 58 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

M. FEWSTER Software Test Automation: Effective Use of Test Execution Tools, Addison
Wesley, 2019
A. KING Programming the Internet of Things: An Introduction to Building Integrated,
Device-to-Cloud IoT Solutions, O'Reilly Media, 2021

G. VENERI, A. CAPASSO Hands-On Industrial Internet of Things: Create a powerful

Industrial IoT infrastructure using Industry 4.0, Packt Publishing, 2018
F. CHANTZIS, I. STAIS, P. CALDERON, E. DEIRMENTZOGLOU, B. WOODS
Practical IoT Hacking: The Definitive Guide to Attacking the Internet of Things, No Starch
Press, 2021
S.J. SHACKELFORD The Internet of Things: What Everyone Needs to Know, Oxford
University Press, 2020
D. MIESSLER The Real Internet of Things, 2017
M. KRANZ Building the Internet of Things: Implement New Business Models, Disrupt
Competitors, Transform Your Industry, Wiley, 2016
H. SMEENK Internet of Things for Smart Buildings: Leverage IoT for smarter insights for
buildings in the new and built environments, Packt Publishing, 2023

Pag. 59 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ONLINE VERSUS IN-PERSON PRODUCTIVITY IN DIFFERENT TEAMWORK

CONTEXTS

Alexia CIUCLEA1
Andreea-Bianca ŞTEFAN2
Ştefan STAN3
Costin Anton BOIANGIU4
Cătălin TUDOSE5

Abstract
This article examines the impact of online and physical environments on productivity
within different team contexts. We analyze many aspects, such as team meetings,
communication methods, and team management, evaluating their influence on
collaboration efficiency.
By identifying key elements, we offer a perspective on the factors that shape productivity
in both digital and traditional environments, contributing to the understanding of the
complexity of current-era collaboration.
Keywords: team productivity, online collaboration, virtual teams, communication
methods, team management, meetings.
JEL Classification: C61

1. Introduction
In today’s age, teamwork is widely accepted as a smart way to organize the operation of an
organization and today’s technology allows such managerial and organizational actions to

1 Student, POLITEHNICA National University for Science and Technology of Bucharest, Romania,
alexia.ciuclea@stud.acs.upb.ro
2 Student, POLITEHNICA National University for Science and Technology of Bucharest, Romania,

andreea.stefan1110@stud.acs.upb.ro
3 Student, POLITEHNICA National University for Science and Technology of Bucharest, Romania,

stefan.stan2402@stud.acs.upb.ro
4 PhD, Professor, POLITEHNICA National University for Science and Technology of Bucharest, Romania,

costin.boiangiu@upb.ro, corresponding author

5 PhD, Lecturer, POLITEHNICA National University for Science and Technology of Bucharest and Luxoft

Romania, Romania, catalin.tudose@gmail.com

Pag. 60 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

be done both physically and digitally. The technological advances of the 90s permitted the
emergence and rise of virtual teams.
At the start of the 21st century, virtual teams have been defined as teams that coordinate
their activities using primarily technological means. Establishing the foundation of
understanding, integration, and trust, putting together the people and the context, involving
the leadership for the knowledge management and information sharing, implementing the
information technology, and acting regarding processes and development – these were steps
that an efficient virtual team needed to cover [1]. Today, these technological means are
represented by instant messaging, video calls, online meetings and conferences, etc.
At their emergence, virtual teams had the purpose of destroying the geographical barriers
present in the conception of teams to be able to select the best individuals from a now much
wider pool of candidates. Virtual teams expand the demographics of the modern workforce
and allow the selection of the best candidates and the improvement of personal flexibility
[2]. Today virtual teams have risen in popularity due to the limitations imposed by a global
pandemic. Therefore, the concept of remote work has settled into the life of the modern
employee. The adoption of teleworking in Europe increased after 2005. Results indicate
that telework has become routine for over 20 percent of all gainfully employed. Expansion
is explained by a working life transition: besides enabling information and communication
technologies, factors associated with managers’ trust and control; the character of jobs,
work tasks, and contracts in knowledge-based industries; and individual and household
work-life balance issues. Telework is connected to permanent employment in the advanced
services sector, slowly diffusing into other sectors. It is increasingly performed in the home
and is becoming more frequent. Individuals with families and children are overrepresented
and among the fastest-growing groups. Broadband access at home is an enabler. Larger
urban regions strengthen their position in favor of teleworking [3].
Although the concept of virtual teams, described and defined two decades ago, is still
relevant, a lot of aspects regarding productivity and communication studied back then are
not applicable anymore. Technology that benefits online collaboration has seen rapid
advancement in the last years and its efficiency proved to be good enough that some modern
online collaboration tools have settled into the physical environment as well. A good
example is instant messaging at the workplace, which has long ago replaced, in most cases,
moving to the colleague’s cubicle or in a meeting room.
This paper will concentrate on how team productivity and efficiency are affected by the
choice of conducting one or more collaboration aspects either digitally or physically.

Pag. 61 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

2. Efficiency of meetings
In some professional contexts, successful teamwork is based on organizing team meetings
to integrate the individual work of the team members and to make sure that everybody is
on the same page moving forward.
Team productivity, in this situation, is represented in two forms: the individual productivity
and performance of team members in the period between the meetings and the efficiency
of communication between the members at the time of the meeting.

2.1. A thorough research

Teleworking has become even more popular after the spread of COVID-19, supported by
rapid technological development. Despite the great convenience, questions about its
productivity arise. Yi Li et al. [4] analyzed the productivity of stock analysts in the context
of online versus offline work environments. The empirical analysis shows that analysts’
forecasts are accurate after offline visits. Additionally, teleworking impacts forecast
accuracy based on analyst experience, resources, firm visibility, and transparency.
Measured in the quality and quantity of forecasts, analysts’ productivity was higher after
participating in traditional meetings rather than in online ones. That is believed to be closely
related to the amount of information gathered from meetings.
Offline meetings consist of face-to-face communication between analysts and that allows
them to engage in a lot more social interaction and build a stronger and closer relationship
with their managers. This enables them to gather sufficient information from the meetings.
Likewise, offline meetings provide an opportunity to visit the company’s headquarters,
which helps with a better understanding of the company’s environment.
Yi Li et al. also discovered that the negative effect of online meetings on analysts’
productivity is more pronounced with analysts with poor ability and limited resources. This
suggests that analysts with enough experience and resources do not rely as much on the
outcome of the meetings and therefore their productivity is not impacted as much by their
quality.
Nonetheless, despite their reduced quality, online activities like annual shareholder
meetings and online board meetings have been found to improve management. Also, in the
context of not being able to hold offline meetings, the online alternative still offers an
information advantage and raises the productivity levels of analysts who have participated
in them besides those who did not.
Overall, while online meetings present some benefits like flexibility and convenience, they
may also have a negative impact on productivity compared to their offline counterparts.

Pag. 62 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

2.2. Online meetings disadvantages

Meetings are ubiquitous at work. Therefore, understanding what makes meetings effective
(or not) is important. Blanchard and McBride [5] have studied the disadvantages of online
meetings amidst the start of the pandemic context. They have found that online meetings
do not allow the same level of engagement as offline meetings do, leading to decreased
motivation and attention of team members. Also, they have discovered that a lack of visual
cues and nonverbal communication can have a negative impact on the effectiveness of
communication between team members during the meeting. When meeting participants
perceive a high enough level of group-ness in their meeting, then they begin to enact the
processes to create a successful meeting and experience the outcomes of a successful
meeting. The authors propose a model connecting the characteristics of successful face-to-
face (FtF) meetings to entitativity and extrapolate this model to online meetings.
We will explain the role that effective communication has in team productivity further in
the article. Additionally, online meetings present technological challenges such as glitches,
poor Internet connection, and software compatibility problems.
Furthermore, online meetings are also at a disadvantage because of the possibility of
experiencing ‘Zoom fatigue’. Fauville et al. [6] have researched Zoom fatigue and found
that people who use video conferencing more frequently, for longer, and with fewer breaks
have expressed more Zoom fatigue. They have also described the disadvantages of Zoom
fatigue: it can lead to exhaustion and burnout among frequent participants, and it can be
amplified or produced by mirror anxiety and hyper-gaze, both of which can distract and
decrease focus. They provided empirical evidence that Zoom fatigue is influenced by the
dynamics of individuals' video-conferencing usage and their psychological experience of
the meeting.
Also, maintaining the effort put into giving nonverbal communication cues required on
video calls drains the participant and reduces the ability to properly understand the
discussion held. On the whole, Fauville et al. have found that Zoom fatigue can harm
productivity by causing the disadvantages discovered, taking away cognitive resources
from actual work tasks.

2.3. Why meetings matter

Even decade-old research acknowledges the importance of meetings and their effect on
productivity. Kauffeld and Lehmann-Willenbrock [7] have found that successful meetings
that contained proactive communication and in which the participants contributed ideas and
suggestions had a positive impact on team productivity. After having recorded 92 regular
team meetings, the study followed the idea that the key to understanding team meeting
effectiveness lies in uncovering the microlevel interaction processes throughout the
meeting.

Pag. 63 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

It has also been found that participating in meetings where they could contribute ideas and
personal goals in an open discussion led to a 10.6% average increase in worker productivity
during a field experiment. The brief participatory meetings also increased treatment
workers’ retention rate and feelings of empowerment such as job satisfaction and a sense of
control [8].
While they have the disadvantages presented above, virtual meetings can also lead to
increased productivity, based on reduced travel time and increased flexibility in scheduling.
During the pandemic, office workers were accustomed to fully remote work or partially
remote hybrid work. Many of them did not intend to go back to commute to a workplace
full-time, no matter the cost. Research from the ADP Research Institute, covering more than
32,000 workers, fully 64% said they would consider looking for a new job if they were
required to return to the office full-time. More than half said they would accept a pay cut of
up to 11% if they could guarantee themselves remote or partially remote hybrid work [9].

3. Communication methods
Amidst the labyrinth of collaborative dynamics, the profound impact of effective
communication on team productivity unfolds as a pivotal dimension that requires scholarly
exploration. This exploration, grounded in the principles of research paper language and
concepts, delves into the intricate facets of communication methods, providing critical
considerations for the optimization of team functioning.
As teams embark on complex missions and objectives, the nuanced exchanges of
information, the articulation of ideas, and the cultivation of mutual understanding emerge
as fundamental determinants shaping the trajectory of collective productivity.
Within the academic discourse on team dynamics, a rigorous analysis of communication
methods is essential to uncover the subtle nuances influencing the efficacy of team
interactions. Scholarly investigations, exemplified by the studies conducted by Li et al. [4],
contribute valuable insights into the impact of exclusive reliance on specific
communication modalities.
This scholarly scrutiny is essential for teams navigating the contemporary landscape, where
the strategic selection of communication methods becomes a decisive factor necessitating
a nuanced understanding of associated advantages and disadvantages.
The examination of communication tools, including email, instant messaging, and verbal
interactions, forms a focal point in this academic exploration. Li et al.'s research [4], for
instance, sheds light on potential delays and varied interpretations resulting from exclusive
reliance on email. Moreover, the comparative studies conducted by Li et al. [4] offer an
objective lens through which the influences of instant messaging and verbal communication
on team productivity can be systematically assessed.

Pag. 64 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Beyond the instrumental considerations of communication methods, the academic inquiry

extends to the profound impact of effective communication among team members on
broader organizational outcomes. Sunarsi et al.'s findings [10] exemplify this, revealing a
correlation between effective team communication and heightened job satisfaction. This
linkage underscores the motivational force embedded within communication quality,
fostering employee engagement and, consequently, contributing intrinsically to the
overarching productivity metrics of a team.
In the ever-evolving landscape of work, this scholarly discourse serves as a compass,
guiding teams toward informed communication strategies aligned with the complex
objectives of contemporary organizational contexts. As teams navigate the intricate
interplay between communication and productivity, these research-driven insights pave the
way for strategic choices that enhance collaborative efficacy and contribute to the enduring
success of teams within dynamic work environments.

3.1. Written Communication Challenges

In scrutinizing written communication, email emerges as a robust tool for documentation;
however, Li et al.'s [4] caution against exclusive reliance on it reveals potential pitfalls.
Relying solely on email can lead to delays and varied interpretations, primarily stemming
from the absence of non-verbal cues.
Moreover, the ubiquity of instant messaging, while fostering real-time interactions,
introduces challenges. It may create an environment pressuring immediate responses,
thereby influencing the overall work pace. A comparative study by Li et al. [4] becomes
imperative to unravel the nuanced influences of both instant messaging and verbal
communication on team productivity.

3.2. Verbal Communication Dynamics

Verbal communication, whether unfolding in physical meetings or through phone calls,
injects a layer of richness into team interactions through non-verbal elements. However, the
transition to the online environment reveals potential challenges.
Technical issues may compromise the advantages brought by gestures and vocal tones,
introducing sound or connection delays. The emphasis from Kauffeld and Lehmann-
Willenbrock's study [11] on proactive communication within meetings underscores the
significance of these sessions beyond information exchange. Such meetings serve as
catalysts for innovation and the expeditious making of crucial decisions.

Pag. 65 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3.3. Team Member Communication Impact

The impact of effective communication among team members, as gleaned from Sunarsi et
al.'s findings [10], extends beyond operational efficiency. It correlates significantly with
higher job satisfaction, fostering motivation and engagement among team members. The
pivotal nature of choosing communication methods becomes evident, guided by
discernment and a comprehensive understanding of their advantages and disadvantages.
This choice is instrumental in cultivating an environment of effective communication,
supported by appropriate methods and practices, ultimately encouraging active
participation and contributing significantly to achieving heightened productivity and
meeting established goals.

3.4 Connectivity in Distributed Teams

Communication emerges as the lifeblood of team unity and awareness. In distributed teams,
maintaining optimal connectivity becomes a prerequisite for efficient collaboration.
However, the online environment, while facilitating remote interactions, poses challenges.
Saghafian and O'Neill's observations [12] shed light on the heightened efforts exerted by
online teams to establish optimal communication channels, underscoring their connectivity
concerns compared to teams operating in physical proximity.
Cang et al. [13] further argue that the absence of physical co-presence and non-verbal
communication contributes to reduced connectivity in online collaborative work. The
delicate balance in connectivity is essential, preventing information overload and
maintaining focus on individual tasks, ultimately preserving and enhancing overall team
productivity.

4. Team management
4.1. Transformational versus transactional
The way a team is led can be an influencing factor in team productivity. Recently, there
have been conducted thorough studies on two leadership styles: transactional and
transformational.
The transactional model is characterized by offering rewards such as bonuses for finishing
clearly defined tasks and by setting punishments to motivate team members. The
transformational model requires a much deeper implication from the team leader in team
dynamics to develop the skills of team members and motivate them by being supportive
and setting high expectations.
Even though both models present very different characteristics, they both contribute to the
growth of job satisfaction. However, the transformational model has been found to have a
Pag. 66 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

stronger impact, suggesting that its motivational aspects have a wider influence on bettering
job satisfaction [10]. Also, this model has proved to be more efficient in the context of
virtual teams because it can be used to strengthen the feeling of trust between team members
[14].
Job satisfaction is a key factor for productivity. The effect culture has on employee
satisfaction is also important. One of the biggest strengths of the organization is the
relationship and communication between the employees and the managers. Research has
discovered that workers with a high level of job satisfaction have a higher chance of being
motivated and engaged in their work-related tasks, which leads to higher productivity [15].

4.2. Challenges of leading teams in an online environment

In an online environment, leaders face new challenges generated by the transition to a
different work setting. Chamakiotis, Panteli, and Davison have found that maintaining
subordinated engagement requires more effort from the leaders [14]. They have also
identified that leaders are required to find new ways to create a feeling of trust between
team members and sustain the work-life balance for both themselves and the team members.
Besides choosing to implement the transformational leadership model, the leader of a
former traditional team must be creative and adapt easily to the numerous challenges that
arise when transitioning to the management of a virtual or partially virtual team.

4.3. Why effective communication matters for leadership

Effective leadership requires physical interaction because it allows a leader to grow strong
interpersonal relationships with team members. Developing strong interpersonal
relationships enhances engagement and satisfaction among team members and leaders who
prioritize it are more likely to inspire and motivate team members, which leads to increased
productivity [16].
As mentioned above, distance takes a toll on effective communication. With the increasing
commonness of remote work, managers must face the challenge of effectively
communicating with team members, which is heavily impacted by time zone differences
[17].

5. Ways of improving productivity

5.1. Using technology to promote effective communication
Yuka, Koyano., et al. [18] claim that one of the main problems of working in an online team
is the meetings because it is difficult to read the intentions and thoughts of the other

Pag. 67 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

participants and the atmosphere of the place. Conventional online meeting systems are
designed on the premise that the participants turn on the camera to show their own faces.
To improve productivity, teams can use avatars that directly reflect their facial expressions.
The system analyzes the participants' facial expressions and reflects them on a cartoon-like
avatar provided.
The program also extends a voting feature that allows participants to vote in speech and
chat. The results of the voting are displayed, and the users can check the differences of
opinion. The system can also show the negative-positive ratio, and the facial expression
ratio of a meeting and generate a word cloud.

5.2. Using the project management method to increase productivity

Turetken et al. [19] identify several ways to increase productivity through Agile project
management. These methods include focusing on the customer (this can help increase
productivity as teams concentrate on delivering products or services that are valuable to
customers); delivering value quickly (helps increase productivity as teams can receive
feedback from customers earlier and make adjustments accordingly); working in small and
autonomous teams (helps increase productivity as teams can communicate and collaborate
more easily); and adapting to changes (helps increase productivity as teams can better cope
with unforeseen changes).
Kirkman et al [20] define team virtuality as the extent to which team members use virtual
tools to coordinate and execute team processes, the amount of informational value provided
by such tools, and the synchronicity of team member virtual interaction. They identified a
key factor that can influence the productivity of a team, the size of the virtual team, and the
implications of their model for management theory and practice.
The dimension that should be the best depends on three main keys: the task characteristics
(the nature of the task can influence how the team uses technology), team characteristics
(their experience with virtual work), and organizational characteristics (the organization's
policies and procedures). These factors still influence the team’s productivity in the present.

5.3. The power of receiving feedback

To enhance online productivity, Rice et al. propose a model for adapting team processes
[21]. 16 teams that worked together for periods ranging from 3 months to 3 years were
studied. Team processes that emerged naturally from long-duration teams were formalized
and taught to shorter-duration teams. These shorter-duration teams comprised 3 different
cohorts, each of which received different levels of training.
This model involves receiving feedback to identify areas that need improvement and
addressing issues; designing new processes tailored to specific needs and objectives;
Pag. 68 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

implementing the new processes and monitoring them; evaluating the effectiveness of the
new processes and adjusting as needed; and maintaining to ensure continuous improvement.
Geister et al. conducted a study with 52 virtual teams formed by students, divided into two
groups: one group received feedback on the tasks they completed, while the other group did
not receive feedback [22]. The experiment's results showed that the group that received
feedback recorded improvements in motivation, satisfaction, and performance. Because
team members were more motivated, they were more satisfied with their team experience
and were able to accomplish more tasks in a shorter time.
The authors of the article suggest that feedback on team tasks had a positive effect through
the following mechanisms: increasing transparency and accountability as team members
better understood how the team functioned and contributed to the development of individual
responsibility; improving communication and collaboration; and increasing trust and
respect among team members. For feedback to have the greatest impact, it needs to be
specific, relevant, timely, and provided by a trustworthy person (e.g., a mentor or team
manager).
Although the effects of feedback were noted in two studies two decades ago, its relevance
remains unmodified. Feedback is a powerful tool to improve the productivity of teams even
nowadays.

6. Discussion
Even though virtual teams were introduced as a revolutionary way to manage a project, that
was two decades ago. Time since then has shown us how hard it is to replicate the quality
of teamwork using virtual means to collaborate.
We have split our research into three categories: meetings, communication, and leadership.
We consider these aspects to be the most important when it comes to the essence of
teamwork. Out of these three, communication is the most important because it stands at the
root of both meetings and leadership.
When it comes to meetings, choosing to stick to physical ones proves to be highly beneficial
due to the use of social interaction. Building strong relationships inside the team is such an
important factor that it influences every aspect of teamwork. Online meetings, however,
present a multitude of challenges that not only disturb the focus of participants but also do
not allow the efficient gathering of information from the other participants.
Leadership is also impacted by the online environment. In comparison to meetings,
leadership doesn’t come with its own specific hardships, and it’s heavily influenced by the
quality of communication. That being said, communication quality is lower in online
settings, and that is reflected in online-based leadership.

Pag. 69 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Lack of visual cues and other non-verbal elements block virtual-based communication from
reaching the nature of traditional communication. Effortlessly staying connected to
teammates is the backbone of collaboration.
Effective communication, motivating leadership, and competent meetings are the
fundamental part of productive and effective teamwork and as such we need to make sure
all three elements are treated vigorously.

7. Further research
The research does not cover the whole subject of the productivity impact of online and
physical environments. While the study provides valuable insight, some aspects remain
unexplored. Mainly, traditional and online meetings were examined, but hybrid meetings
have become more relevant. Similarly, the study focused on leadership in fully remote and
fully in-person meetings, while hybrid leadership contains elements from both styles. The
research showed the importance of project management to increase team productivity,
however, details of a comparison between online and in-person methods were not primarily
focused on and would provide valuable insight.

References
[1] C.B. GIBSON, S.G. COHEN (Eds.) Virtual teams that work: Creating conditions for
virtual team effectiveness, 2003 Hoboken, NJ: John Wiley & Sons
[2] A.M. TOWNSEND, S. M. DEMARIE, A.R. HENDRICKSON Virtual teams:
Technology and the workplace of the future. Academy of Management Executive, 1998, 12:
17–29.
[3] B. VILHELMSON, E. THULIN Who and Where are the Flexible Workers?, Exploring
the Current Diffusion of Telework in Sweden, New Technology, Work and Employment
2016, 31, 1, 77–96.
[4] Yi LI, Wei ZHANG, Pengfei WANG Working online or offline: Which is more effective?
Research in International Business and Finance, Volume 58, 2021, 101456, ISSN 0275-
5319
[5] A. L. BLANCHARD, A. MCBRIDE Putting the “Group” in Group Meetings:
Entitativity in Face-to-Face and Online Meetings Managing Meetings in Organizations,
2020, 71–92.
[6] F. GÉRALDINE, L. MUFAN, A. QUEIROZ., N. JEREMY, J. BAILENSON., J.
HANCOCK Video-conferencing usage dynamics and nonverbal mechanisms exacerbate
Zoom Fatigue, particularly for women Computers in human behavior reports, 2023,
10:100271-100271.
Pag. 70 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[7] S. KAUFFELD, N. LEHMANN-WILLENBROCK Meetings Matter. Small Group

Research, 2011, 43(2), 130–158.
[8] J. DOMÍNGUEZ Having a voice in your group: Increasing productivity through group
participation, 2022, 1-21.
[9] L. KUGLER The impact of virtual meetings Communications of The ACM, 2022,
65(11):19-21.
[10] I. SUNARSI, I.G. PUTRA, I. N.W. SUWARJANA The effect of effective
communication on employee job satisfaction and organizational commitment. International
Journal of Human Resource Studies, 2021, 11(2), 117-132.
[11] S. KAUFFELD, N. LEHMANN-WILLENBROCK Proactive communication in
meetings: Antecedents, outcomes, and moderators Academy of Management Journal, 2011,
54(5), 1022-1040.
[12] M. SAGHAFIAN, T.E. O'NEILL Connectivity and information sharing in distributed
teams: A case study of information technology project teams International Journal of Project
Management, 2018, 36(10), 1370-1385.
[13] S. CANG, L. WANG, Z. CHEN, B. WU The impact of physical co-presence and non-
verbal communication on online collaborative work: An empirical study. Information and
Management, 2023, 64, 103224.
[14] P. CHAMAKIOTIS, N. PANTELI, R.M. DAVISON Reimagining e-leadership for
reconfigured virtual teams due to COVID-19. Int J Inf Manage. 2021 Oct;60:102381.
[15] B.J. ALI, G. ANWAR An Empirical Study of Employees’ Motivation and its Influence
Job Satisfaction. International Journal of Engineering, Business and Management, 2021,
5(2), 21-30.
[16] S. MÜLLER The Challenge of Leadership Development in the Digital Era, 2022, 94-
106.
[17] T. WETTER Managing Teams in the Digital Age, 2023, 31-44.
[18] K. YUKA, O. TOSHIAKI, A. SHINGO, D. SAYAKA, M. HIROYUKI, Y.
TAKAYUKI, T. HIROAKI Development and Evaluation of Online Meeting System to
Promote Effective Communication, 2021, 712-725.
[19] O. TURETKEN, A. JAIN, A.A. YAYLA, L.M. MOLINA A review and analysis of
literature on agile project management. International Journal of Production Economics,
2016, 159, 79-90.
[20] B.L. KIRKMAN, J.E. MATHIEU The dimensions and antecedents of team virtuality
Journal of Management, 2005, 31(5), 700-718

Pag. 71 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[21] D.J. RICE, B.D. DAVIDSON, J.F. DANNENHOFFER Improving the Effectiveness of
Virtual Teams by Adapting Team Processes. Comput Supported Coop Work, 2007, 16, 567–
594 (2007).
[22] S. GEISTER, U. KONRADT, G. HERTEL Effects of process feedback on motivation,
satisfaction, and performance in virtual teams Small Group Research, 2006, 37(5), 459–
489

Bibliography
L. WAIZENEGGER, R. ULRICH, R. MAIER. Connectivity Changes: Why Do Project
Team Members Shift between Different States of Connectivity? 2014
S. MARZIEH, K. O'NEILL A Phenomenological Study of Teamwork in Online and Face-
to-Face Student Teams Higher Education, 2018, 75(1):57-73
L. XI, R. CANG., D. GUERRA, P. MCLAREN, J. VYAS, K. REKIMOTO, E. MACLEAN
Demonstrating Virtual Teamwork with Synchrobots: A Robot-Mediated Approach to
Improving Connectedness, 2023
CENTER FOR THE ADVANCEMENT OF VIRTUAL ORGANIZATIONS Winning in the
Virtual Workplace: 10 Experts Reveal How to Lead Your Remote Team to Success, 2024
SOBIA PUBLICATION VIRTUAL TEAMS Design Your Successful Remote Team Culture:
Managing a Remote Team as a Leader. Virtual Team Management & Motivation, 2023
H. OSMAN Influencing Virtual Teams: 17 Tactics That Get Things Done with Your Remote
Employees, 2016
M. SCANNELL Big Book of Virtual Teambuilding Games: Quick, Effective Activities to
Build Communication, Trust, and Collaboration from Anywhere!, 2011
R. LEPSINGER, D. DEROSA Virtual Team Success: A Practical Guide for Working and
Leading from a Distance, 2010
S. M. ALAM Effective communication in virtual teams: Strategies for Seamless
Collaboration in Remote Work Environments, 2024
K. DEGNER Guide to Virtual Team Building, 2021
T.S. HOLLEMA Virtual Teams Across Cultures: Create Successful Teams Around the
World, 2020
K. OZENC, G. FAJARDO Rituals for Virtual Meetings: Creative Ways to Engage People
and Strengthen Relationships, 2021
S. M. BRAGG Managing Virtual Teams, 2024

Pag. 72 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

K. PITCH Microsoft Teams Guide for Success: Mastering Communication, Collaboration,

and Virtual Meetings with Colleagues & Clients, 2024
K. HALL, A. HALL, N. PATTERSON Leading Remote and Virtual Teams: Managing
Yourself and Others in Remote and Hybrid Teams or When Working from Home, 2021
W. GARCIA Communication Challenges in Global Virtual Teams: Leading People around
the World, 2021
M. FRAIDENBURG Mastering Online Meetings: 52 Tips to Engage Your Audience and
Get the Best Out of Your Virtual Meetings, 2020
P. BISHOP Remote Team Success: Strategies for Building High-Performing Virtual Teams:
The Ultimate Remote Working Playbook, 2024
K. FISHER, M. FISHER Manager's Guide to Virtual Teams, 2011
M.R. LEE Leading Virtual Project Teams: Adapting Leadership Theories and
Communications Techniques to 21st Century Organizations, 2021
C. LEMA Building & Managing Virtual Teams: Five Ways to Create a High-Performance
Culture for Remote Workers, 2012
O. IKE The Art of Remote Collaboration: Strategies for Effective Teamwork in a Virtual
World, 2024
Y. ZOFI A Manager's Guide to Virtual Teams, 2011
SOBIA PUBLICATION Virtual Teams - Design Your Successful Remote Team Culture:
Managing a Remote Team as a Leader. Virtual Team Management & Motivation, 2023
J. CHEN Engaging Virtual Meetings: Openers, Games, and Activities for Communication,
Morale, and Trust, 2020
P. F. ALEXANDER, M. HULL, P. VAN DER HAGEN The 7 Habits of Highly Effective
Virtual Teams: Make a Success of Your Virtual Global Workforce, 2014
T. HIRATA The Virtual Leader: How to Manage a Remote Workplace, 2022

Pag. 73 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

FACILITATING ITALIAN LANGUAGE ACQUISITION AMONG ROMANIAN

SPEAKERS IN AN ONLINE SETTING: THE ADVANTAGES OF INTEGRATING
DIGITAL RESOURCES

Mariana COANCĂ1

Abstract
This study examines the integration of digital tools and flipped classroom strategies in the
online delivery of an A1-level Italian course for Romanian speakers. Using an online course
from a well-known platform, instruction combined multimedia input, Romanian-language
scaffolding, and cross-linguistic comparisons to support early language acquisition.
Supplementary resources such as collaborative activities and gamified platforms enhanced
vocabulary development, grammatical accuracy, and learner engagement. Feedback
indicated that flexible preparation, multimodal materials, and immediate corrective support
contributed significantly to the learners’ progress, with English proficiency aiding access
to external resources. The findings highlight the effectiveness of a digitally enriched,
learner-centered design for beginner-level Italian language instruction in online education.
Keywords: online course, cross-linguistic comparisons, digital pedagogy, multimedia
input, feedback
JEL Classification: I21, I23

1. Introduction
Italian language is widely recognized for its melodic and expressive qualities, characterized
by phonetic clarity and rhythmic intonation, making it appealing to a diverse audience,
including business professionals, travelers, and language enthusiasts. While Standard
Italian is promoted for official and educational purposes, there is a growing movement to
preserve regional languages and dialects such as Neapolitan, Sicilian, Venetian, Lombard,
and Sardinian, as integral components of Italy’s cultural heritage [1]; [2].
Globally, Italian ranks among the top 25 most spoken languages, with approximately 67
million native speakers in the European Union and up to 85 million speakers worldwide
[3]; [4]. Although its global reach may not match that of English or Mandarin, Italian has
significant value due to its rich artistic, literary, musical, and architectural heritage.

1Associate Professor, PhD., Department of Foreign Languages, Romanian-American University, 1B

Expozitiei Blvd., sector 1, Bucharest, Romania, mariana.coanca@rau.ro
Pag. 74 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Proficiency in Italian provides us with direct access to these cultural treasures in their
original form.
Beyond cultural attractions, Italian offers strategic advantages in various professional
fields. Italy’s global influence in fashion, automotive design, gastronomy, and the fine arts
enhances the utility of Italian in international markets. Consequently, individuals are often
motivated to learn Italian due to personal relationships, such as family ties or romantic
partnerships, as well as professional collaborations.
From a cognitive perspective, bilingualism has been associated with enhanced cognitive
reserve, potentially delaying the onset of age-related cognitive decline and conditions like
dementia [5]; [6]. Linguistically, Italian shares significant structural and lexical similarities
with other Romance languages, including French, Spanish, Portuguese, and Romanian.
This typological proximity facilitates cross-linguistic transfer, making Italian more
approachable for students who already speak another Romance language, as they can utilize
familiar grammatical patterns and cognates.

2. The Acquisition of Italian by Romanian Speakers

Third language (L3) acquisition is a well-established domain within the field of

multilingualism. It investigates the processes by which individuals who already possess
proficiency in two or more languages acquire an additional language and examines how
their existing linguistic repertoire influences this learning trajectory. The acquisition of
Italian as a third language involves a dynamic interplay of linguistic proximity, cognitive
processes, and pedagogical practices. This section draws on research in language transfer,
language proximity, cross-linguistic transfer, and the role of technology in facilitating
language acquisition.
Romanian and Italian share a high degree of structural, lexical, and phonological similarity,
which lowers the learning barrier for native Romanian speakers engaging with Italian as a
third language. One of the most prominent features aiding mutual intelligibility is the
abundance of cognates (i.e. words that are etymologically related and often retain similar
meanings and forms across both languages). Examples such as mamă (Romanian) and
mamma (Italian), or casă (Romanian) and casa (Italian) facilitate intuitive vocabulary
recognition for students. These linguistic parallels often allow Romanian speakers to infer
the meanings of unfamiliar Italian words with minimal effort. Furthermore, both languages
exhibit similar verb conjugation systems, including regular and irregular patterns, as well
as subject-verb-object sentence structure, for example, the Romanian sentence “Ea vorbește
limba italiană” and its equivalent sentence in Italian “Lei parla l’Italiano”. While no Italian
dialect is directly close to Romanian, there are notable linguistic similarities between
Romanian and certain southern and central Italian dialects particularly in vocabulary,
grammar, and phonetics due to their shared Latin roots [7]; [8].

Pag. 75 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Syntactic parallels such as noun-adjective agreement and gendered articles provide students
with a familiar grammatical framework. For instance, constructions like un câine frumos in
Romanian and un cane bello in Italian reflect almost identical structural logic, making them
easily transferable. However, one important grammatical difference lies in the gender
system: while Italian operates with two grammatical genders (masculine and feminine),
Romanian maintains three (masculine, feminine, and neuter). This asymmetry may lead to
confusion, particularly when Romanian neuter or feminine nouns correspond to masculine
nouns in Italian, potentially interfering with article and adjective agreement during
production (“libro” in Italian is a masculine noun whereas its Romanian equivalent “carte”
is a feminine noun; “giorno” in Italian is a masculine noun whereas its Romanian equivalent
“zi” is a feminine noun).
As for the pronunciation of words, Romanian and Italian both maintain five core vowel
sounds with relatively stable pronunciations, and their consonant systems and stress
patterns follow predictable and comparable rules. This phonetic proximity greatly aids
Romanian speakers in acquiring accurate Italian pronunciation and prosody. However, as
[9] points out, second language learners often face challenges in phonology, vocabulary,
and grammar due to interference from their first language (L1) habits. Consequently,
according to [10], the comprehensible input hypothesis posits that language acquisition is
most effective when students are exposed to input that is both understandable and slightly
beyond their current level of proficiency. A multimodal strategy that incorporates such input
across diverse, authentic contexts promotes repeated encounters with key language
structures, thereby supporting both understanding and practical use.
[11] provides insights into the dynamics of L3 acquisition by examining the roles of the
first language (L1) and second language (L2) in L3 production. His findings reveal that
learners often rely on their L2, especially when it shares similarities with the L3. However,
when acquiring a third language, Romanian speakers may experience language interference
by transferring linguistic elements from their native language to the Italian language. A
common example is the phonological interference, such as the pronunciation of the letter
h. In Italian, h is silent, while in Romanian it is clearly articulated. As a result, Romanian
speakers at A1 level of Italian proficiency may overpronounce h in Italian words like ho,
hai, and ha (forms of the verb essere), reflecting transfer from their native phonological
system.
Other researchers seek to understand the broader conditions under which linguistic transfer
occurs. Specifically, in a systematic review they examine what factors determine when,
how, and to what extent previous linguistic experience, whether from the first language,
second language, or both, affects the initial stages and beyond of adult third language (L3)
acquisition [12]. Finally, [13] advocate for linguistically responsive teaching practices in
multilingual settings. Their work emphasizes the need for educators to recognize and
leverage learners’ existing linguistic resources, promoting a more inclusive and effective
language learning environment.

Pag. 76 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Recognizing the role of sociocultural engagement in virtual settings, the integration of

digital tools and platforms expands opportunities for meaningful interaction and
personalized learning, enhancing both the accessibility and effectiveness of online Italian
instruction. Digital platforms have revolutionized language learning by offering flexible,
learner-centered approaches. Tools such as Duolingo, and Coursera provide interactive
content, immediate feedback, and self-paced progression, which are valid elements that
support effective acquisition of Italian. Among freely available tools for Italian language
learning, the course “Parliamo in Italiano” on Edulia Treccani – Scuola2 stands out for its
comprehensive and methodologically sound structure. Developed by one of Italy’s leading
linguistic institutions, Treccani, the course is designed for adult and young adult learners
acquiring Italian as a foreign language, and spans four levels aligned with the Common
European Framework of Reference for Languages (CEFR), covering A1–A2 and B1–B2
proficiency levels [14].

3. Instructional Design for A1-Level Italian: Integrating Digital Tools in an Online

Setting

This section outlines the instructional design and teaching strategies employed in the online
delivery (via Microsoft Teams) of an A1-level Italian course for Romanian speakers, with
particular attention to the integration of external digital tools to enhance third language
acquisition. The course, Parliamo in Italiano, was used to teach fifteen Romanian learners
whose second language is English. The primary objective was to guide learners toward
successfully completing the A1 level of Italian proficiency within three months.
To ensure the successful implementation of online resources, it was crucial to consider a
well-prepared instructional design and the digital competence of both the teacher and the
learners. For the teacher, digital competence involved not only the technical ability to
navigate and use various educational technologies but also the pedagogical knowledge to
integrate these tools meaningfully into the learning process. This included selecting
appropriate resources, adapting content for online delivery, and providing ongoing support
for them in using these tools. For a learner, digital competence meant having the skills to
engage with the online course materials, participating in digital activities, troubleshooting
common technical issues, and collaborating with peers through online platforms.
The online course supports both guided instruction and independent study, making it
suitable for classroom integration as well as autonomous learning [15]. What distinguishes
“Parliamo in Italiano” is its integration of multimedia content, interactive exercises, and
authentic spoken input. Each thematic unit is centered around communicative situations
drawn from everyday life, beginning with a video sequence that features authentic spoken

2 Source: https://inclasse.edulia.it/resources?type=parliamoitaliano

Pag. 77 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Italian and is available with or without subtitles [16]. These video inputs are intended to
spark learner motivation while activating cognitive and perceptual mechanisms essential to
language acquisition [17]. Morphosyntactic, pragmatic, conversational, and lexical features
of Italian are contextualized within these scenarios and further explored through a range of
focused tasks within each unit [18], as shown below.

Figure 1: Screenshot of the overview of activities in Unit 2, Lesson 1 from the

“Parliamo in Italiano” course, Edulia Treccani - Scuola3

Figure 2: Screenshot of a Lesson Interface from “Parliamo in Italiano” course, Edulia

Treccani - Scuola: “Ti va di andare al cinema?”4

3 Source: https://inclasse.edulia.it/resources?menu%5B_tags%5D=parliamoitaliano
4 Source: https://inclasse.edulia.it/resources?menu%5B_tags%5D=parliamoitaliano
Pag. 78 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

To achieve the objectives of an instructional design, the teacher selected the lesson “Ti va
di andare al cinema?” from the “Parliamo in Italiano” course (see Figure 2). This lesson
introduces foundational communicative competencies relevant to everyday social
interactions, including making and refusing invitations, expressing availability, and
employing modal verbs such as “potere”, “volere”, and “dovere”.

The Flipped Classroom Model

In recent years, the flipped classroom model has gained significant attention in language
education as a means of enhancing learner engagement, autonomy, and active classroom
participation. The flipped model, as defined by [19], enhances adult learners’ engagement
by shifting initial input and rule discovery to individual preparation, allowing class time to
focus on communicative practice and collaborative tasks.
In the context of Italian as a foreign language (IFL), the model has been applied across
educational levels, including beginner (A1) courses, with promising results. For example,
a foundational study examining the integration of flipped learning in a university-level
beginner Italian course, emphasizes that the flipped model, when aligned with the principles
of inductive learning and communicative pedagogy, can enhance learner autonomy and in-
class collaboration through a cyclical structure of pre-class video content, interactive class
practice, and post-class reinforcement [20].
Adding to recent innovations, Carmelo Galati’s “Gratis!: A Flipped-Classroom and Active
Learning Approach to Italian” offers an open-access flipped classroom resource specifically
designed for Italian language instruction [21]. The textbook integrates asynchronous
learning materials such as video lectures and digital exercises with active, student-centered
classroom activities aimed at developing communicative competence. Expanding the
application of technology-enhanced flipped models, Laura Capitani’s blended learning
course at Maastricht University offers another example of successful integration. As
reported by [22], Capitani combines Babbel’s online modules with supplementary
resources and individualized Skype sessions: students first engage independently with
structured Babbel lessons, then reinforce their skills through additional practice activities,
before participating in one-to-one tutorials focused on communicative practice. This model
not only promotes learner autonomy and flexibility but also personalizes instruction,
allowing students to progress at their own pace and receive immediate, targeted feedback
[23].
Finally, in their study on the broader adoption of flipped learning in Italian schools, [24]
report that educators employ the model to foster student-centered learning environments.
Their findings suggest that technology accessibility and teacher training are pivotal factors
influencing the model’s success. Although their study does not focus solely on Italian
language instruction, it reflects the growing institutional support for flipped pedagogies in
Italy.

Pag. 79 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Implementation
Prior to the online session, learners engaged with the video of the lesson and accompanying
comprehension activities. At the beginner stage, they are still developing a basic
understanding of the new language. To support this process, the teacher initially explained
key lesson concepts in Romanian, ensuring that students understood essential vocabulary,
grammatical rules, and sentence structures related to making invitations, expressing
refusals, and offering alternatives. This approach allowed students to grasp the material
more easily, drawing on their familiarity with Romanian linguistic structures. For example,
when teaching the conjugation of verbs in the present tense, the teacher highlighted the
similarities between the two languages, helping students recognize the shared patterns and
structure. Drawing explicit cross-linguistic comparisons during instruction can enhance
comprehension, reduce cognitive load, and promote positive transfer from Romanian to
Italian in the early stages of communicative language use. A comparison between the two
languages reveals important similarities that can facilitate the acquisition of these
competencies. Both Italian and Romanian, as Romance languages, use specific verbs to
express ability (potere / a putea), necessity (dovere / a trebui), and desire (volere / a vrea).
However, while Italian typically combines these modal verbs directly with an infinitive
(e.g., Posso leggere il tuo nome – “I can read your name”), Romanian often introduces an
additional particle (să) before the infinitive (e.g., Trebuie să citesc mesajul – “I must read
the message”). Understanding these structural differences can help them transfer existing
knowledge while adapting to the syntactic patterns of Italian.
While the course provides structured and comprehensible content, its implementation in an
online environment could be significantly enhanced by incorporating supplementary digital
tools and platforms. For instance, the video of the lesson was replayed during class with
strategic pauses to facilitate clarification and pronunciation drills. Learners were also
introduced to supplementary resources, such as “The Beginner Podcast series” from “News
in Slow Italian”5, which is specifically designed for A1 level of proficiency. This podcast
delivers news content at a slower pace, using simplified language to support language
acquisition [25]. To reinforce their understanding of modal verbs such as dovere, potere,
and volere, the teacher incorporated an interactive digital grammar quiz focused on
practicing these modal verbs, which are essential for expressing intentions and managing
invitations. In this exercise (see Figure 3), they completed sentences by selecting the correct
conjugated verb form based on context such as Perché Francesco non ___ venire alla festa?
(potere) or Noi ___ vedere quel bosco (volere)6.
These structured yet contextually rich exercises provided immediate, scaffolded practice
with high-frequency verbs crucial for A1-level communicative competence. The visually
clear interface and focused input helped support the development of grammatical accuracy

5 Source: https://www.newsinslowitalian.com/
6 Source: https://www.newsinslowitalian.com/series/grammar/27/186/quiz/2

Pag. 80 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

while maintaining learner engagement. Although the platforms support a limited selection
of languages, excluding Romanian, they were able to successfully engage with the tasks
thanks to their intermediate English proficiency (ranging from B1 to B2), minimizing
potential language barriers. They are increasingly accustomed to navigating various digital
platforms for diverse purposes and their familiarity with online environments likely
facilitated their engagement with the tasks and enabled them to manage the technological
aspects of the learning process with greater ease.

Figure 3: Screenshot of a grammar activity from “News in Slow Italian”, The Beginner
Podcast series, practicing modal verbs7

Social Interaction
To enhance social interaction, the teacher adopted an innovative approach using Microsoft
Teams. Learners were paired in breakout rooms to role-play invitations and refusals in real-
time. One illustrative A1-level breakout activity, titled “Inviting a Friend Out”, focused on
the use of functional expressions like Ti va di...?; Che ne dici di...? together with the modal
verbs (potere, volere, dovere). To model the task, the teacher performed a brief sample
dialogue with a student: Che ne dici di andare al cinema sabato? / Mi dispiace, non posso.
Vado a trovare mia madre. / Allora, facciamo domenica? / Va bene, volentieri!. Key
expressions were shared via screen or chat to support comprehension. Students, working in
pairs, were then instructed to co-construct a short dialogue including one invitation, one
polite refusal with a reason, and one accepted alternative, using at least three of the target
structures. The activity followed a structured timing breakdown: 2 minutes to enter breakout
rooms, 3 minutes to review instructions, 8 minutes to write and rehearse the dialogue, 5
minutes to perform and self-correct, and 2 minutes to return to the main room. This

7 Source: https://www.newsinslowitalian.com/series/grammar/27/186/quiz/2

Pag. 81 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

scaffolded role-play not only reinforced functional language use but also promoted peer
collaboration and real-time negotiation of meaning in a communicative context. In
alignment with the sociocultural theory and the Interaction Hypothesis, these dialogic
exchanges played a crucial role in facilitating the internalization of syntactic structures and
pragmatic norms, particularly among adult learners capable of metalinguistic reflection
[26]; [27].
As part of a personalization-oriented technique, students were tasked with writing
customized dialogues in which they invited peers to events of personal interest (e.g.,
concerts, dinners, sports, etc.). These dialogues were composed and shared collaboratively
via the Files section of Microsoft Teams. For instance, one pair created a conversation in
which a student invited a classmate to a live concert using the expression “Ti va di venire
al concerto sabato sera?”, followed by a polite refusal and an alternative proposal. Peers
were encouraged to view, comment on, and suggest edits directly within the shared
document, fostering a sense of co-authorship and reflective learning.
In parallel, the teacher provided targeted written feedback through the document’s comment
feature, praising accurate usage of expressions (e.g., Benissimo! Ottimo uso di “Ti va di...”)
and gently correcting errors related to syntax or verb conjugation. In some instances,
teachers added reformulations or guiding questions (e.g., Potresti aggiungere
un’alternativa più specifica?) to promote revision and metalinguistic awareness. This
scaffolded feedback loop enhanced the pedagogical value of the activity by supporting both
accuracy and learner autonomy. [28] emphasizes that learners engage more deeply with
materials that connect to their social identities and experiences. In this context, the
opportunity for personalization combined with collaborative authorship and timely
feedback, made the language practice both meaningful and memorable.
Finally, to reinforce vocabulary acquisition and grammatical structures introduced during
instruction, as well as to sustain learner motivation in the online environment, the teacher
the teacher developed interactive exercises to review the introduced vocabulary and
structures. For example, they were tasked with matching expressions to appropriate
responses (e.g., Ti va di uscire? → Mi dispiace, non posso), completing sentence gaps using
modal verbs (e.g., Voglio andare al cinema, ma non posso, devo studiare), and selecting
context-appropriate invitations from multiple-choice options, such as:

Ce înseamnă “Ti va di andare al cinema?”

A) Vrei să mergi la cinema?
B) Mergi la cinema?
C) Poți să mergi la cinema?
D) Nu vreau să merg la cinema.

Care este răspunsul corect la invitația “Ti va di mangiare fuori stasera?” dacă nu poți
merge?
A) Mi dispiace, non posso.

Pag. 82 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

B) Va bene, ci vediamo alle 7.

C) Non voglio.
D) Sì, mi piace mangiare fuori.

Care dintre următoarele este un refuz politicos la o invitație?

A) Sì, con piacere!
B) Mi dispiace, non posso.
C) Andiamo!
D) Mi piace molto l’idea.

These questions helped them connect vocabulary and expressions to their real-world
equivalents in Romanian, reinforcing comprehension and recall.

4. Insights and Perspectives

Requesting feedback allows instructors to ensure that their teaching aligns with learners’
needs and to adjust accordingly. In this study, the teacher determined that the most practical
moment to collect feedback was at the end of the lesson, using the Microsoft Teams chat
function. At this point, all had already engaged with the content and were better positioned
to reflect on the effectiveness of the session. The teacher prepared questions to allow them
to reflect on different aspects of their learning experience, from content preparation to
engagement and interaction. Key questions included: How effective did you find the online
resources in helping you prepare for the lesson?; How did the online format impact your
participation?; What part of the lesson did you find most engaging or useful? Why?; Were
the multimedia resources helpful in reinforcing the lesson content? If yes, how? and What
would you recommend for progressing to the next level of Italian proficiency?
They actively shared their insights, and the teacher observed that, in an online environment,
they tended to provide feedback more readily than students participating in face-to-face
settings. To illustrate this, they emphasized that access to pre-class video materials and
multimedia resources enabled them to prepare at their own pace, fostering a sense of
autonomy and boosting their confidence during live sessions. This flexible preparatory
phase supported deeper engagement with the content and promoted mental readiness for
interactive classroom activities.
Moreover, they found the combination of visual and auditory input particularly beneficial
for vocabulary retention and comprehension, as it allowed for more effective internalization
of new structures. In-class pauses during video playback were especially appreciated as
they offered moments to reflect on the language, identify grammatical patterns, and
consolidate their understanding of modal verbs and functional expressions through guided
scaffolding.
Finally, the integration of external resources that give access to grammar tutorials and
contextualized video examples was also perceived as highly supportive. Most of these tools
Pag. 83 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

were accessed in English, due to the limited availability of Romanian-language materials

on mainstream platforms. Collaborative in-class activities, including role-plays, dialogue
construction, and peer interaction in breakout rooms (via Microsoft Teams), were
considered meaningful opportunities for communicative practice and personalization of the
language. The use of collaborative digital tools further encouraged creative expression and
enabled them to receive written feedback, which they regarded as particularly useful for
developing grammatical accuracy and reflective learning strategies. Gamified platforms
were also highlighted as effective tools for enhancing engagement. These tools encouraged
active recall in a low-pressure environment, and the immediacy of feedback they provided
was especially valued for allowing timely self-correction and reinforcement of vocabulary
and grammar.
Although the course was delivered online and entirely in Italian, posing occasional
challenges, they felt that the availability of support materials in English effectively
compensated for the absence of Romanian-language resources. Their ability to draw on
English-language explanations and examples was identified as a key factor in their
progress. They also observed that their mother tongue, Romanian, positively influenced
their acquisition of Italian, particularly in the development of speaking skills. They
identified structural and lexical similarities between the two Romance languages, which
facilitated pronunciation, the intuitive use of cognates, and the construction of basic
sentence patterns. However, several learners expressed a preference for a blended learning
approach, suggesting that a combination of online and in-person instruction could enhance
the learning experience. Such a format would foster stronger social connections and
encourage more spontaneous spoken interaction.

5. Conclusion
The Italian language has both cultural significance and professional value, particularly in
globally influential sectors such as tourism, fashion, design, and gastronomy. For Romanian
speakers, Italian acquisition is notably facilitated by the structural, lexical, and phonological
similarities between the two Romance languages, which support accelerated
comprehension and production through the recognition of cognates and comparable
grammatical patterns.
The design and delivery of the A1 lesson, conducted online, via Microsoft Teams,
demonstrated the pedagogical effectiveness of digitally mediated instruction. The
integration of external digital resources, such as “News in Slow Italian” and gamified
learning platforms, further enriched the multimodal, learner-centered environment,
promoting scaffolded development and fostering autonomy. Learners consistently
emphasized the benefits of combining visual, auditory, and interactive content, which
contributed to improved vocabulary retention, grammatical understanding, and
pronunciation accuracy. Although the absence of Romanian-language support materials
Pag. 84 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

presented some challenges, participants successfully utilized their English proficiency as a

mediating tool in their third language (L3) acquisition.
Future adaptations of the course could benefit from a blended learning model, combining
online flexibility with in-person sessions to enhance interpersonal communication,
strengthen group cohesion, and boost learner motivation through spontaneous interaction.
In conclusion, the instructional model employed in this study offers a flexible and effective
framework for beginner-level Italian learning in online contexts, leveraging digital tools
and learner-centered strategies to support communicative competence in third language
acquisition.

References
[1] M. Maiden and C. Parry, Linguistic Diversity in Italy: The Impact of Regional
Languages on the National Identity. Cambridge: Cambridge University Press, 2018.

[2] S. Tufi and J. Blackwood, The Dialects of Italy: Linguistic Variation and Cultural
Identity. London: Routledge, 2019.

[3] WorldoStats, “Italian Language - WorldoStats.” [Online]. Available:

https://worldostats.com/demographic/languages-by-number-of-
speakers/#google_vignette. [Accessed: 22-Apr-2025].

[4] Wikipedia, “Italian Language - Wikipedia.” [Online]. Available:

https://en.wikipedia.org/wiki/Italian_language. [Accessed: 22-Apr-2025].

[5] T. Bak, The Cognitive Benefits of Bilingualism. Cham: Springer, 2015.

[6] D. Perani and F. Abutalebi, The Cognitive Effects of Bilingualism: Implications for
Learning and Memory. Oxford: Oxford University Press, 2018.

[7] R. Posner, The Romance Languages. Cambridge: Cambridge University Press, 1996.

[8] M. Sala, De la latină la română: Evoluția istorică a limbii române într-un context
comparativ romanic. Bucharest: Univers Enciclopedic, 1998.

[9] H. B. Beardsmore, Bilingualism: Basic Principles. Avon: Tieto, 1982.

[10] S. Krashen, The Input Hypothesis: Issues and Implications. London: Longman, 1985.

Pag. 85 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[11] B. Hammarberg, “Roles of L1 and L2 in L3 Production and Acquisition,” in Cross-

Linguistic Influence in Third Language Acquisition: Psycholinguistic Perspectives, J.
Cenoz, B. Hufeisen, and U. Jessner, Eds. Bristol: Multilingual Matters, 2001, pp. 21–41.

[12] E. Puig-Mayenco, J. González Alonso, and J. Rothman, “A systematic review of

transfer studies in third language acquisition,” Second Language Research, vol. 36, no. 1,
pp. 31–64, 2018.

[13] A. Krulatz, G. Neokleous, and A. Dahl, Multilingual Approaches to Additional

Language Teaching: Bridging Theory and Practice. Bristol: Multilingual Matters, 2022.

[14] Treccani, “Parliamo in Italiano – Course Info,” Edulia Treccani - Scuola, 2016.
[Online]. Available: https://inclasse.edulia.it/resources?type=parliamoitaliano. [Accessed:
26-Apr-2025].

[15] Treccani, “Parliamo in Italiano – Course Info,” Edulia Treccani - Scuola, 2016.
[Online]. Available: https://inclasse.edulia.it/resources?type=parliamoitaliano. [Accessed:
26-Apr-2025].

[16] Treccani, “Parliamo in Italiano – Course Info,” Edulia Treccani - Scuola, 2016.
[Online]. Available: https://inclasse.edulia.it/resources?type=parliamoitaliano. [Accessed:
26-Apr-2025].

[17] Treccani, “Parliamo in Italiano – Course Info,” Edulia Treccani - Scuola, 2016.
[Online]. Available: https://inclasse.edulia.it/resources?type=parliamoitaliano. [Accessed:
26-Apr-2025].

[18] Treccani, “Parliamo in Italiano – Course Info,” Edulia Treccani - Scuola, 2016.
[Online]. Available: https://inclasse.edulia.it/resources?type=parliamoitaliano. [Accessed:
26-Apr-2025].

[19] J. Bergmann and A. Sams, Flip Your Classroom: Reach Every Student in Every
Class Every Day. Washington, DC: International Society for Technology in Education,
2012.

[20] M. Caruso and F. Verdina, “Flipping the Classroom in Beginner Italian: Implications
for Teaching and Learning,” in Enabling Learning: Language Teaching for Australian
Universities, J. Kinder, N. Fraschini, and M. Caruso, Eds. Canberra: ANU Press, 2024,
pp. 125–156.

Pag. 86 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[21] C. Galati, Gratis!: A Flipped-Classroom and Active Learning Approach to Italian.

Philadelphia: North Broad Press, 2022.

[22] G. Pedrotti, “Learning Italian at Maastricht University with a Blended Course and
Babbel,” Babbel Magazine, 21 Oct. 2019. [Online]. Available:
https://www.babbel.com/en/magazine/learning-italian-at-maastricht-university-with-
babbel. [Accessed: 26-Apr-2025].

[23] G. Pedrotti, “Learning Italian at Maastricht University with a Blended Course and
Babbel,” Babbel Magazine, 21 Oct. 2019. [Online]. Available:
https://www.babbel.com/en/magazine/learning-italian-at-maastricht-university-with-
babbel. [Accessed: 26-Apr-2025].

[24] G. Pieri and L. Laici, “L’approccio Flipped Classroom nel Movimento ‘Avanguardie
Educative’,” Italian Journal of Educational Technology, vol. 25, no. 3, pp. 55–67, 2017.

[25] Linguistica 360, “News in Slow Italian – Beginner Podcast Series,” News in Slow
Italian. [Online]. Available: www.newsinslowitalian.com/beginners. [Accessed: 26-Apr-
2025].

[26] L. S. Vygotsky, Mind in Society: The Development of Higher Psychological

Processes. Cambridge, MA: Harvard University Press, 1978.

[27] M. H. Long, “Native Speaker/Non-Native Speaker Conversation and the Negotiation

of Comprehensible Input,” Applied Linguistics, vol. 4, no. 2, pp. 126–141, 1983.

[28] Z. Dörnyei, The Psychology of Second Language Acquisition. Oxford: Oxford

University Press, 2009.

Bibliography

L. Cirillo and N. M. Doerr, Eds., Italian as a Foreign Language: Teaching and

Acquisition in Higher Education. Vernon Press, 2020.

R. Foote, “Transfer in L3 Acquisition: The Role of Typology,” in Third Language

Acquisition and Universal Grammar, Y. I. Leung, Ed. Bristol: Multilingual Matters,
2009, pp. 89–114.

Pag. 87 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

R. B. Kunz, Comparative Grammar of Spanish, Portuguese, Italian, French, Romanian

and Catalan, 2nd ed. Charleston, SC: CreateSpace Independent Publishing, 2015.

D. Singleton and L. Ryan, Language Acquisition: The Age Factor, 2nd ed. Bristol:
Multilingual Matters, 2004.

I. Tyrou, “Teaching Italian as a Foreign Language and Elements of Its Culture Using
Digital Media,” ResearchGate, 2023. [Online]. Available:
https://www.researchgate.net/publication/373103308.ResearchGate+1Re. [Accessed: 26-
Apr-2025].

B. Woll and L. Wei, Cognitive Benefits of Language Learning: Broadening our

Perspectives. The British Academy, 2019.

A. Zingaro, "Forliviamo: How to Incidentally Learn Italian Through a Context-Aware

Mobile Application," in Technological Innovation Put to the Service of Language
Learning, Translation and Interpreting: Insights from Academic and Professional
Contexts, Ó. Ferreiro-Vázquez and A. T. V. Moutinho Pereira Correia, Eds. Berlin: Peter
Lang, 2023, pp. 143–156.

Italiano Bello, “Che ne dici di: Meet Up with Your Friends in Italian.” [Online].
Available: https://www.italianobello.it/che-ne-dici-di. [Accessed: 20-Apr-2025].

Dacorti, “Italian Dialects: Exploring Italy’s Rich Linguistic Diversity,” Attractive Italy, 7
Jan. 2025. [Online]. Available: https://attractiveitaly.com/italian-dialects-exploring-italys-
rich-linguistic-diversity/. [Accessed: 27-Apr-2025].

Pag. 88 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

EXPLORING THE LINK BETWEEN INNOVATION AND ICT READINESS. A

COMPARATIVE ANALYSIS OF ESTONIA AND ROMANIA

Daniela Alexandra CRIȘAN1

Justina Lavinia STĂNICĂ2

Abstract
This paper examines the relationship between the Summary Innovation Index (SII) and
certain ICT readiness indicators, including the Network Readiness Index (NRI) and the
Human Resources in Science and Technology (HRST) Index, across the EU-27 countries.
The objective is to gain a deeper understanding of how digital infrastructure and human
capital contribute to innovation performance within the region. The results obtained
confirm the previous research of the authors of this paper regarding the close connection
between innovation and ICT.
In the second part of the paper, we conducted a case study in which we analyzed the
innovation and ICT readiness landscapes in Estonia and Romania, two countries that
present a clear contrast.
Keywords: ICT, innovation performance, correlation, Summary Innovation Index (SII),
Network Readiness Index (NRI), Human Resources Science and Technology (HRST) Index
JEL Classification: M15

1. Introduction

The European Union has long recognized that fostering innovation is key to driving
economic growth, enhancing global competitiveness, and addressing societal challenges.
Therefore, at the EU level, innovation-driven growth has become a key priority, focusing
on directing resources to foster research and development. Investments in research,
development, education, and skills are crucial policy areas for the EU, as they play a vital
role in driving economic growth and establishing a knowledge-based economy.

This paper examines the relationship between different indicators of innovation and ICT
readiness. Recent studies have highlighted the importance of aligning innovation strategies

1Associate Professor, PhD, School of Computer Science for Business Management, Romanian-American
University, e-mail: daniela.alexandra.crisan@rau.ro, corresponding author
2Lecturer, PhD, School of Computer Science for Business Management, Romanian-American University,

lavini.stanica@rau.ro
Pag. 89 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

with ICT readiness frameworks across European nations [1]. Also, the role of digital
transformation in enhancing innovation performance is analyzed in [2]. In addition,
previous research conducted by the authors shown a strong linkage between innovation and
ICT, considering various indicators, such are: the Summary Innovation Index (SII) and the
the Global Innovation Index (GII), for the innovation, and, respectively the Networked
Readiness Index (NRI), the ICT Development Index (IDI), and various EUROSTAT
indicators, for ICT [3][4]. Therefore, by analyzing the models provided, we aim to better
understand the demand for and supply of highly qualified individuals in science and
technology, alongside the essential need for continued investments in R&D.

Materials and methods

For the materials and methodology, all charts and graphics included in this paper were
created by the authors using Microsoft Excel (including ANOVA analysis), with data
sourced from the European Innovation Scoreboard (EIS) 2024, the Global Information
Technology Report 2024, and Eurostat. The HRST indicator values were updated on 13
December 2024.

The Summary Innovation Index (SII)

The European Innovation Scoreboard (EIS) is an initiative launched by the European

Commission to provide a comparative analysis of innovation performance across European
countries, both at national and regional levels. EIS is published annually, and it measures
and ranks countries based on their innovation capabilities and outcomes, helping to identify
trends and gaps in innovation across Europe.

The purpose of the EIS is to assess different aspects of innovation, through a variety of
indicators:

1. Human Resources – the availability of skilled individuals in science and

technology.
2. Research Systems – the performance of research institutions and public research
expenditure.
3. Innovation-Related Investments – private sector investments in innovation and
R&D.
4. Enterprise and Innovation Activities – the presence and performance of
innovative companies and startups.
5. Intellectual Assets – indicators like patents, trademarks, and designs.
6. Market and Economic Impact – the contribution of innovation to economic
performance, including productivity and job creation.

The initiative came to life in 2001, but over time, it has undergone changes. At the
beginning, the report was published under the name "Innovation Union initiative". The
current name of EIS was acquired in 2015. In addition, the indicators in the early versions
Pag. 90 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

of the EIS focused on more traditional aspects of innovation, such as R&D investment,
patents, and scientific publications. Their structure has changed over, in order to reflect
more accurately the changes of the economies. One of the most recent, important updates
of EIS happen in 2021, they were meant to better reflect evolving trends in innovation.
Some of the key changes introduced in 2021 include: (1) Expanding Indicators on Digital
Transformation, (2) Introduction of the "Green Innovation" Dimension, (3) New
Methodology for Composite Indicator.

In recent editions, the methodological framework of the scoreboard is based on a total of

32 indicators, divided into four main categories and 12 dimensions, to assess the innovation
performance of the EU, its Member States and selected third countries (EIS, 2024). Based
on these indicators, a composite index, Summary Innovation Index (SII) is calculated
yearly for every country:

• 1.1 Human resources • 2.1 Finance and support

• 1.2 Attractive research • 2.2 Firm investments
systems • 2.3 Use of information
• 1.3 Digitalisation tehnologies

FRAMEWORK
INVESTMENTS
CONDITIONS

INNOVATION
IMPACTS
ACTIVITIES
• 3.1 Innovators
• 3.2 Linkages • 4.1 Employment impacts
• 3.3. Intellectual assets • 4.2 Sales Impacts
• 4.3 Enviromental stability

Figure 1. The structure of the Summary Innovation Index (SII) in 2024

(chart made with Ms. Excel, by authors)

Based on their SII index values, the 27 Member States are categorized into four
performance groups (Figure 2):

• Innovation leaders (Denmark, Sweden, Finland, and the Netherlands) – perform in

innovation well above the EU average (above 125% than EU average);
• Strong Innovators (Belgium, Austria, Ireland, Luxembourg, Germany, Cyprus,
Estonia, and France) – innovate below the leaders, but above the EU average;
• Moderate innovators (Slovenia, Spain, Czechia, Italy, Malta, Lithuania, Portugal,
Greece and Hungary) – perform in innovation below or equal to the EU
average (at least 70% of EU average);
• Emerging innovators (Croatia, Poland, Slovakia, Latvia, Bulgaria and Romania) –
innovate well below the EU average.
Pag. 91 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Summary Innovation Index in 2024

200
Summary Innovation Index

150
100
50
(SII)

Greece

Spain
Latvia

EU-27
France

Austria

Denmark
Romania

Czechia

Finland
Italy
Bulgaria

Poland
Croatia
Hungary

Slovenia

Cyprus

Ireland

Netherlands
Slovakia

Malta

Estonia

Germany
Luxembourg
Portugal
Lithuania

Belgium

Sweden
Figure 2: The four groups of EU-27 countries by their innovation performance (SII)
relative to EU scores, in 2024
(chart made with Ms. Excel, by authors)

We note that, according to EIS 2024: “All performance scores described in this report are
relative to that of the EU in 2017 to facilitate the tracking of progress and trends that enable
policymakers to identify specific areas requiring attention through strategies and
programmes at national level.” [5, pg. 9].

Measuring the ICT readiness

In order to assess the ICT readiness of a country, we have used two indicators:
1. HRST Index measures the availability and quality of human capital in science,
technology, engineering, and mathematics (STEM) fields.
2. NRI assesses a country’s overall readiness to leverage information and
communication technologies (ICT) for economic and social development.

The Human Resources in Science and Technology (HRST) Index is a key measure used
by Eurostat to assess the availability of skilled labour in science and technology fields
across European countries. It includes people with higher education qualifications in
science and technology fields, such as those with degrees in engineering, life sciences,
physical sciences, and information technology [6].

In our analysis, we have used the HRST indicator with the following dimensions:

• Unit of measure: Percentage of population in the labour force (PC_ACT)

• Category: Persons with tertiary education (ISCED) and/or employed in science and
technology (HRST)
• Age class: From 15 to 24 years and 65 to 74 years (Y15-24_Y65-74)

Using the HRST Index, we can rank the 27 Member States as follows:
Pag. 92 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 3: The EU-27 countries ranked by their S&T performance (HRST Index 2023)
(chart made with Ms. Excel, by authors)

The Network Readiness Index (NRI) is also a composite index meant to offer a
description of the digital economy of a country. It is published as part of the Global
Information Technology Report [7], as a result of collaboration between various partners,
including the World Economic Forum (WEF) and INSEAD. The 2024 version was
published under the motto “Building a Digital Tomorrow: Public-Private Investments and
Global Collaboration for Digital Readiness”, by the Portulans Institute in collaboration with
Saïd Business School, University of Oxford.

The NRI is built around four main pillars, each of which is broken down into sub-pillars
and indicators:

Pag. 93 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• ICT Infrastructure • Skills

• Technology • Usage
Adoption

TECHNOLOGY
PEOPLE
(Infrastructure
(Skills and
and
Usage)
availability)

IMPACT GOVERNANCE
(Outcomes (Policies and
and Results) Regulation)
• Economic Impact • Regulatory
• Social Impact Environment
• Government and
Business
Coordination

Figure 4. The structure of the Network Readiness Index (NRI) in 2024

(chart made with Ms. Excel, by authors)

In the next graphic the ranking of the 27 EU members, according to their NRI values. The
four pillars Technology, People, Governance and Impact are also represented:

The Network Readiness Index (NRI) and its pillars

for the EU-27 members in 2024
100
90
80
70
NRI

60
50
40
30
Romania

Austria

France

Finland
Latvia

Denmark
Greece

Hungary

Poland

Italy
Spain

Germany
Croatia

Bulgaria

Cyprus

Ireland
Slovakia

Slovenia

Netherlands
Malta

Portugal

Luxembourg
Belgium
Lithuania

Czechia

Estonia

Sweden

NRI Technology People Governance Impact

Figure 5. The Network Readiness Index (NRI) and its pillars for the EU-27 members in
2024
(chart made with Ms. Excel, by authors)

Pag. 94 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

2. Correlations between innovation and ICT for the EU-27 Member States

In this section, we aim to examine the correlation between the SII and ICT readiness
indicators, such as the NRI and HRST Index, for the EU countries in order to better
understand how digital infrastructure and human capital influence innovation performance
across the region.

2.1 Summary Innovation Index (SII) vs Network Readiness Index (NRI)

The next diagram illustrates the Summary Innovation Index (SII) and Network Readiness
Index (NRI) indexes for the 27 EU Member States in 2024 suggesting that a linkage exists
between them. This visual representation indicates that countries with higher NRI scores
tend to have stronger innovation performance, as reflected by their SII:

Comparative analysis of SII and NRI

for UE-27 Members in 2024
80
Summary Innovation Index (SII)

150
75
130
70
110
65 90
60 70
55 50
50 30

Denmark
Slovenia

Luxembourg
Ireland

Belgium
Romania

Latvia

Poland

Hungary

Italy

France
Estonia

Finland
Bulgaria

Croatia

Greece

Cyprus

Austria

Netherlands
Spain

Germany
Slovakia

Malta
Portugal
Lithuania

Czechia

Sweden

NRI SII

Figure 6. Comparative analysis of SII and NRI for UE-27 Members in 2024
(chart made with Ms. Excel, by authors)

Our next step was to describe this relationship using a linear model. The figure below shows
that the two composite indexes for 2024 are linearly correlated:

Pag. 95 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 7. Correlation between SII and NRI for UE-27 Members in 2024
(chart made with Ms. Excel, by authors)

The equation of the linear regression is provided:

SII = 3.50 x NRI - 118.19

R = 0.86

further demonstrating the strength of this correlation and offering a quantitative basis for
understanding the connection between human capital in science and technology and
innovation performance across EU Member States.

2.1 Summary Innovation Index (SII) vs. the Human Resources in Science and
Technology (HRST) Index

Our next investigation referred to the relationship between the Summary Innovation Index
(SII) and the Human Resources in Science and Technology (HRST) Index for the EU-27,
at the global level, for the period 2017 to 2024:

Pag. 96 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 8. Correlation between SII and HRST Index at EU-27 level in 2024
(chart made with Ms. Excel, by authors)

As expected, they are also strongly correlated because both indicators reflect key drivers of
innovation. The HRST Index measures the availability of a skilled workforce in science
and technology, which is essential for fostering research, development, and technological
advancements.

3. Comparative analysis between Romania and Estonia in terms of innovation and

ITC readiness

Romania faces significant challenges in terms of innovation, which have hindered its
progress in the global digital economy. Despite having a relatively young and educated
workforce, the country struggles with inadequate infrastructure and limited access to high-
speed internet, particularly in rural areas. This digital divide hampers the widespread
adoption of new technologies and stifles the growth of digital businesses. Moreover, while
the country has made strides in some sectors, the public and private sectors are often slow
to embrace digital transformation, with outdated governance structures and regulatory
frameworks that do not fully support innovation.

Additionally, Romania’s education system, though improving, still falls short in providing
the necessary digital skills to equip the workforce for the demands of the modern economy.
As a result, Romania’s innovation ecosystem remains underdeveloped, with limited
investment in research and development, low rates of collaboration between businesses and
academia, and insufficient support for startups and digital entrepreneurship.

Pag. 97 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

These factors contribute to Romania’s relatively low ranking in global innovation indices,
hindering its competitiveness in an increasingly digital world.

In contrast, Estonia stands out as a leader in innovation. Estonia’s regulatory frameworks

are designed to encourage investment in research and development, collaboration between
businesses and academic institutions, and the growth of digital entrepreneurship.

As a result, Estonia consistently ranks among the top performers in global innovation and
ICT readiness indices, contrasting sharply with Romania's situation:

Comparative analysis of SII

for Estonia and Romania in the period 2017-2024
130
Summary Innovation Index (SII)

110
90
70
50
30
10
-10 2017 2018 2019 2020 2021 2022 2023 2024

Romania Estonia EU-27

Comparative analysis between NRI and sub-pillars

for Estonia and Romania in 2024

NRI

Technology

People

Governance

Impact

0 10 20 30 40 50 60 70 80 90

Estonia Romania

Pag. 98 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 9. Comparative analyses using SII, NRI and HRST Indicators for Estonia and
Romania
(chart made with Ms. Excel, by authors)

Examining the relationship between innovation and the Human Resources in Science and
Technology (HRST) Index, for both Estonia and Romania, we have found a strong link,
that suggests that the availability and quality of skilled labor in science and technology
plays a crucial role in fostering innovation in both countries:

Pag. 99 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 10. Correlation between SII and NRI for Estonia and Romania (bottom) in the
period 2017-2024
(chart made with Ms. Excel, by authors)

For Estonia, this strong connection indicates that the country’s investment in developing a
highly skilled workforce in science, technology, engineering, and mathematics (STEM) has
directly contributed to its innovation success. Estonia’s high HRST score reflects a well-
educated, digitally literate population that is prepared to drive innovation in emerging
technologies.

Romania’s lower HRST score suggests that there are still gaps in terms of the quantity and
quality of its skilled workforce in STEM fields. Despite having a relatively high level of
education, Romania faces challenges in producing enough highly skilled individuals with
specialized knowledge in science and technology, which could be stifling its innovation
potential. A lack of investment in education and research, coupled with an underdeveloped
R&D sector, limits the country’s ability to generate and implement innovative solutions.
Improving the HRST index by investing in STEM education, skills development, and
research infrastructure could significantly boost Romania's innovation capacity, helping it
better compete in the global digital economy.

In summary, the strong relationship between HRST and innovation in both countries
underlines the importance of a highly skilled workforce in driving technological
advancements and fostering a vibrant innovation ecosystem. Estonia’s success shows how
a strong HRST index can be a major driver of innovation, while Romania’s situation
suggests that enhancing its HRST capacity is key to unlocking its innovation potential.

Pag. 100 / 350

Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

4. Conclusions
In conclusion, the analysis of the correlation between the Summary Innovation Index (SII),
Network Readiness Index (NRI), and the Human Resources in Science and Technology
(HRST) Index reveals a clear and significant relationship between digital infrastructure,
human capital, and innovation performance across the EU countries. The findings show that
countries with higher HRST and NRI scores tend to perform better in innovation,
suggesting that investments in digital readiness and a skilled workforce are crucial factors
for fostering innovation.

Romania and Estonia are at different poles in terms of innovation, digitalization and
innovation. The analyses carried out reinforce the idea that a highly skilled workforce and
a strong digital infrastructure led to better performance in innovation, emphasizing the need
for strategic investments in education, research, and digitalization in Romania’s case, in
order to bridge the gap with leading innovators in the EU.

Acknowledgment: The research was carried out in part within the Center for
Computational Science and Machine Intelligence (CSMI) of the Romanian-American
University's School of Computer Science for Business Management".

References

[1] Sfetcu N. (2024), “ICT in the EU: Driving Digital Transformation and Innovation,” IT
& C, 4:1, pp. 41-57, DOI: 10.58679/IT96537, available at:
https://www.internetmobile.ro/ict-in-the-eu-driving-digital-transformation-and-
innovation/
[2] Varzaru A.A., Bocean C.G. (2024), “Digital Transformation and Innovation: The
Influence of Digital Technologies on Turnover from Innovation Activities and Types of
Innovation,” MDPI Open Access Journals, available at: https://www.mdpi.com/2079-
8954/12/9/359
[3] Crișan D.A., Stănică J.L. (2022), “Analysis of the correlation between Innovation and
ICT Readiness for the EU-27 Member States,” Journal of Information Systems and
Operations Management (JISOM), Vol. 16 No. 2 / Dec 2022, pp. 35-45
[4] Preda A.M., Crișan D. A., Stănică J.L. (2014), “The Impact of ICT on Innovation
Performance in Europe. Case of Romania,” Journal of Information Systems and
Operations Management (JISOM), Vol 8 No 1, 2014, pp. 1-12, Ed. Universitara
[5] European Innovation Scoreboard 2024, available at:
https://op.europa.eu/en/publication-detail/-/publication/8a4a4a1f-3e68-11ef-ab8f-
01aa75ed71a1/language-en
Pag. 101 / 350
Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[6] EuroSTAT, available at: https://ec.europa.eu/eurostat/statistics-

explained/index.php?title=Glossary:Human_resources_in_science_and_technology_(HRS
T)
[7] Global Information Technology Report, available at: https://kpmg.com/xx/en/our-
insights/transformation/kpmg-global-tech-report-2024.html

Bibliography
[1] Crișan D.A., Stănică J.L. (2022), “Analysis of the correlation between Innovation and
ICT Readiness for the EU-27 Member States,” Journal of Information Systems and
Operations Management (JISOM), Vol. 16 No. 2 / Dec 2022, pp. 35-45
[2] European Innovation Scoreboard 2024 Report, available at:
https://op.europa.eu/en/publication-detail/-/publication/8a4a4a1f-3e68-11ef-ab8f-
01aa75ed71a1/language-en
[3] Global Information Technology Report, available at: https://kpmg.com/xx/en/our-
insights/transformation/kpmg-global-tech-report-2024.html
[4] Preda A.M., Crișan D. A., Stănică J.L. (2014), “The Impact of ICT on Innovation
Performance in Europe. Case of Romania,” Journal of Information Systems and
Operations Management (JISOM), Vol 8 No 1, 2014, pp. 1-12, Ed. Universitara
[5] Sfetcu N. (2024), “ICT in the EU: Driving Digital Transformation and Innovation,” IT
& C, 4:1, pp. 41-57, DOI: 10.58679/IT96537, available at:
https://www.internetmobile.ro/ict-in-the-eu-driving-digital-transformation-and-
innovation/
[6] Varzaru A.A., Bocean C.G. (2024), “Digital Transformation and Innovation: The
Influence of Digital Technologies on Turnover from Innovation Activities and Types of
Innovation,” MDPI Open Access Journals, available at: https://www.mdpi.com/2079-
8954/12/9/359
[7] EuroSTAT, available at: https://ec.europa.eu/eurostat/statistics-
explained/index.php?title=Glossary:Human_resources_in_science_and_technology_(HRS
T)

Pag. 102 / 350

Article’s total number of pages: 14
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

USABILITY TESTING APPLIED TO GRAPHICAL AND VR APPLICATIONS IN

AN ACADEMIC ENVIRONMENT

Crina DUTA1
Nicoleta Luminita CĂRUȚAȘU2
George CĂRUȚAȘU3
Ionuț-Cristian PREDERIC4

Abstract
This article explores usability testing in the context of graphic and virtual reality (VR)
applications in academic settings, emphasizing its importance for optimizing user
experience. Usability testing assesses how easily and effectively users can interact with an
application, being crucial for improving the design and functionality of digital products.
The study compares user perceptions of the usability of graphical applications versus VR
applications, using data collected through a questionnaire addressed to students. The results
reveal significant differences in how users perceive ease of use, satisfaction, and efficiency
between the two types of applications, highlighting the unique challenges and potential of
VR applications in education.
Keywords: Usability testing, Virtual reality, User management systems, Technology-
based improvement
JEL Classification: M15

1. Introduction
In recent years, graphics applications and virtual reality (VR) technology have become
increasingly present in various fields, from entertainment and education to medicine and
engineering. These technologies provide immersive interactive experiences, allowing users
to explore virtual environments and interact with digital elements in a more intuitive and
natural way. In the academic environment, graphics applications and VR are used for

1PhD student, Eng., National University of Science and Technology POLITEHNICA Bucharest,
crina_andreea.duta@upb.ro, corresponding author
2Prof. PhD Nicoleta Luminita Carutasu (University POLITEHNICA of Bucharest, UPB),

nicoleta.carutasu@upb.ro
3Prof. PhD George Carutasu (Romanian-American University, URA), george.carutasu@rau.ro
4PhD student, Eng., National University of Science and Technology POLITEHNICA Bucharest,

ionut.perederic@upb.ro
Pag. 103 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

simulations, complex visualizations and hands-on training, contributing to more effective

and engaging learning.
A key aspect in the development of these applications is usability testing, which evaluates
how easily and efficiently users can interact with a system. Good usability ensures a smooth
experience, reducing user frustration and optimizing the learning process or their
productivity. Especially in the case of virtual reality, where interaction differs significantly
from that with traditional applications, usability testing becomes crucial for identifying
specific challenges and improving interface design. By comparing user perceptions of the
usability of classic and VR graphics applications, valuable insights can be gained into how
these technologies can be optimized to meet user needs.
As technology evolves, user expectations regarding the interface and functionality of digital
applications are becoming increasingly high. A well-designed interface must be intuitive,
efficient, and accessible, reducing the time required for learning and adaptation. Especially
in the case of virtual reality, where users interact through gestures, movement, and
specialized controllers, the experience can vary considerably depending on the interface
design and the level of comfort offered. Therefore, usability testing plays a key role in
optimizing the user experience, contributing to the development of more intuitive and
efficient applications, both in academia and in other fields.

2. Graphics applications and VR applications

Graphics applications are software designed for creating, editing, and manipulating visual
elements, such as images, illustrations, 3D models, or digital interfaces. They play an
essential role in a wide range of fields, facilitating design, simulation, and visualization
processes [1].
One of the main areas of application is the academic one, where tools like MATLAB,
Blender or EndNote Visual are used for modeling, graphical visualization and data
organization. MATLAB is recognized for its data processing and simulation capabilities,
which are frequently used in engineering and exact sciences. On the other hand, Blender is
an open-source solution for 3D modeling and animation, used in education and research for
creating complex visual models [2].
In the medical field, graphics applications are essential for medical imaging analysis and
anatomical simulations. 3D Slicer and Mimics are relevant examples of software that allow
the visualization of anatomical structures and facilitate computer-assisted surgical
interventions. These tools contribute significantly to the training of healthcare professionals
and the accurate diagnosis of patients [3].
The entertainment industry also relies heavily on graphics applications for creating visual
content. Unity, Adobe Photoshop, and Illustrator are among the most widely used programs
for game development, image editing, and graphic design. Unity is a game engine used for
developing interactive experiences, including virtual and augmented reality. Adobe
Pag. 104 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Photoshop and Illustrator are standards in graphics editing, being widely used by designers
and visual artists [4].
Therefore, graphic applications represent an essential tool in multiple fields, contributing
to the improvement of the creative process, research and education. The continuous
evolution of these technologies allows the development of increasingly efficient solutions,
adapted to the needs of users [5].
Virtual reality is a technology that creates a three-dimensional digital simulation of an
environment, allowing the user to interact with this space in a more realistic way. This
experience is made possible by equipment such as VR headsets, motion controllers, and
sometimes even body sensors. The main goal of VR is immersion, that is, the feeling that
the user is physically present in the virtual world, although they are in a real space [6].
One of the most valuable uses of virtual reality is in education. For example, students can
explore the human body in detail in an interactive way or visit archaeological sites without
leaving the classroom. VR is also used in professional training, such as training pilots,
doctors, or workers in hazardous environments, by providing controlled and safe scenarios
for practice [7].
In medicine, virtual reality is used for both therapeutic and educational purposes. VR
exposure therapy is applied in the treatment of disorders such as phobia, PTSD or anxiety.
Surgeons can also simulate interventions before performing them, improving the accuracy
and safety of medical procedures [8].
Video games were among the first industries to adopt VR on a large scale. Through this
technology, gamers have access to interactive experiences in which body movements
influence the action in the game. In addition, VR is used in cinema, where viewers can
experience interactive films, becoming part of the story.
In addition to education, health and entertainment, VR is also making its way into
architecture, tourism, industrial design and psychology. For example, architects can present
interactive 3D models of buildings to clients before they are built. In tourism, visitors can
virtually explore tourist destinations around the world. These applications increase
efficiency and creativity in design and decision-making processes.

3. Usability testing
Testing, in a broad sense, is the process of verifying whether a product, system, or service
functions according to established requirements. This can occur in any field – from
education, where testing means assessing knowledge, to the software industry, where
testing involves identifying errors and validating the functionality of applications.
Essentially, testing plays an essential role in ensuring quality, reducing risks, and increasing
trust in the final product. It is performed either manually, by people who interact directly
with the product, or automatically, with the help of software tools that simulate user
behavior. Testing is not just about finding defects, but also about validating that the product
Pag. 105 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

meets the real needs of users, that it is safe, efficient, and easy to use. A good practice is for
testing to be integrated throughout the development of a product, not just at the end, to
prevent major problems before they become costly or difficult to fix.
Testing is of several types through functional testing, performance testing, security testing,
compatibility testing, regression testing, automated testing and usability testing.
Functional testing checks whether the application or system does exactly what it is supposed
to do, according to specifications. Buttons, forms, functions, etc. are tested. Performance
testing evaluates how well the system behaves under road, whether it is fast, stable and
efficient when used intensively or by many users. Security testing deals with identifying
vulnerabilities, to prevent unauthorized access or data loss.
Compatibility testing checks whether the product works correctly on different devices,
operating systems, browsers or versions. Regression testing ensures that recent changes
(bug fixes or new features) have not broken things that worked well before. Automated
testing is performed using programs or scripts that run tests automatically, streamlining the
process.
Usability testing is the process of evaluating how easily and effectively a user can interact
with a product, system, or digital interface. The goal of this testing is not only to identify
design errors, but also to understand the behavior and expectations of real users. By
observing how they navigate and perform specific tasks, designers and developers can
adjust the product so that the experience is intuitive and enjoyable. Testing is usually done
with a small group of users but strategically selected to cover different types of use.
This method is essential in the development of applications, websites or software, as it
highlights accessibility and functionality issues early on that could affect end-user
satisfaction. More than a technical check, usability testing involves empathy – that is,
putting yourself in the user’s shoes to understand what obstacles they might encounter. The
results are then used to refine the product and create solutions that are better suited to the
real needs of the target audience.
Usability testing begins with establishing clear objectives: what exactly you want to
evaluate in terms of user experience. Typically, you choose a few essential tasks that users
need to complete, such as filling out a form, finding information, or placing an order. It is
important that the scenarios reflect real-world situations so that the test results are relevant.
Before the test, participants are recruited who are representatives of the target audience, i.e.
people who might be among the future users of the product.
Testing sessions take place in a controlled setting, either physical or online, where
participants are asked to navigate the product and express their thoughts out loud. During
this time, the team of observers notes what actions the users take, where they get confused,
what mistakes they repeat, and how they react emotionally. It is important that the testers
do not intervene or guide the participant but only observe. Tools that record the screen,
mouse movements, or gaze (eye-tracking) can also be used for more detailed analysis.
Pag. 106 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

After the tests are completed, the data is centralized and analyzed to extract useful
conclusions. Patterns are looked for: frequently encountered problems, frustrations, or
unintuitive steps. Based on this information, recommendations are made to improve the
design or interaction flow. Usability testing is usually not done just once but is repeated
after each major release of the product, to ensure the most fluid and satisfying experience
for end users.
Usability testing brings multiple benefits, starting with the early identification of issues that
would otherwise negatively affect the user experience. By directly observing how people
interact with a product, the development team can better understand the real needs and
expectations of users, not just what was assumed during the design phase. This
understanding leads to concrete improvements to the interface, which makes the product
more intuitive, more efficient, and more enjoyable to use. In addition, usability testing helps
reduce costs in the long run, because problems solved early are much cheaper than those
corrected after launch. At the same time, a well-optimized product from a usability
perspective increases user satisfaction, loyalty, and even conversions, in the case of
commercial sites.

4. The comparative approach – Research method

The main object of this research is to test the usability of graphic and virtual reality
applications in the educational context, with a focus on improving the user experience. In
an era where technology is becoming increasingly present in the learning process, it is
essential to understand how these applications are perceived and used by pupils and
students. Usability, defined as the ease with which a user can interact efficiently and
satisfactorily with a system, is an essential criterion for the success of any interactive
application. In the case of educational applications based on virtual reality, good usability
can significantly contribute to active learning, user involvement and information retention.
Thus, evaluating the user experience is not just a technical stage, but a crucial step for
optimizing the educational process through innovative digital means.
To analyze students' perceptions of the usability of these applications, the questionnaire
method was used - known in the specialized literature as the survey method. This research
method is quantitative and involves the systematic collection of data through a standardized
instrument, namely the questionnaire, which includes closed and/or open questions. The
questionnaire is an effective method of obtaining information from a large number of
respondents in a relatively short time, while also offering the possibility of analyzing the
answers in an objective and comparable way.
The questionnaire method is frequently used in social sciences, including educational
research, because it allows for direct insights from real users. In the present context, this
tool provided a clear picture of how students interact with graphical and VR applications,
what difficulties they encounter, what elements they find useful, and how they perceive the
impact of these technologies on their own learning. Thus, the data obtained through this

Pag. 107 / 350

Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

method contributes to a deeper understanding of how technology can be adapted and

improved to better support the educational process.
The questionnaire method is frequently used in social sciences, including educational
research, because it allows for direct insights from real users. In the present context, this
tool provided a clear picture of how students interact with graphical and VR applications,
what difficulties they encounter, what elements they find useful, and how they perceive the
impact of these technologies on their own learning. Thus, the data obtained through this
method contributes to a deeper understanding of how technology can be adapted and
improved to better support the educational process.
The other three sections focused on direct user experience. The second part tracked how the
participant interacted with the traditional graphical application, while the third part
analyzed the interaction in the VR environment. The questions were designed to capture
the perceived level of difficulty, the degree of involvement, but also the preferences of the
users. The last section, focused on learning and improvement, targeted how the participants
felt they had progressed or adapted during the use of the applications. Together, these four
components help us understand not only which application is more effective, but also why
it is perceived that way and for whom, providing a complete picture of the user experience.
A total of 70 students were involved in this research, who voluntarily completed the
questionnaire developed to evaluate the usability of graphic and virtual reality applications
in the academic field. The diversity of those who responded is important because the
perception of technology, the degree of familiarity with new digital tools and the way of
interacting with them can vary depending on gender, which can influence the way the
usability of the analyzed applications is perceived. The sample was made up of students
from various study programs, most of them enrolled in fields with a technological or
creative component, such as computer science, multimedia, graphic design or digital
communication, fields in which the use of graphic and VR applications has an increasing
practical relevance.
It is essential to mention that although all 70 students fully completed the general section
of the questionnaire, only 44 of them also responded to the part dedicated to virtual reality
applications. The other 26 respondents stated that they had no previous contact with such
applications and, therefore, could not provide relevant feedback in this regard. This aspect
highlights an important reality regarding the degree of penetration of VR technology in the
academic environment – namely, that, despite its educational potential, virtual reality is not
yet widely or uniformly used among students. The lack of direct experience with VR
applications limits the ability to effectively assess their usability, but, at the same time,
highlights the opportunity to expand access to these technologies for educational purposes.
Thus, the results obtained from the 44 respondents who actually interacted with VR
applications will be analyzed separately, in order to provide a clearer picture of their
experience and perceptions in relation to the usability of these innovative tools. Below in
Figure 1 we see the distribution of those who responded according to their gender.

Pag. 108 / 350

Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 1. Gender of the respondents

The analysis of the demographic and academic profile of the respondents reveals that the
majority of the students who participated in this research are in their third year of
undergraduate studies. They are followed, in number, by fourth-year students and, to a
lesser extent, by students enrolled in master's programs. This structure reflects an increased
interest in the field of graphic applications and emerging technologies, especially among
students in advanced stages of their academic career, when contact with complex
technological applications becomes more frequent and more applied within the study
programs.
Regarding the field of study, approximately 80% of respondents are enrolled in bachelor's
or master's programs in Computer Science, while the remaining 20% come from the field
of Electronics. This distribution confirms the increased relevance of the researched topic
for students in technical and scientific fields, especially for those who are already familiar
with advanced notions of programming, digital interfaces and interactive technologies, such
as those used in virtual reality.

Figure 2. Year of study

Figure 3. Field of study

Pag. 109 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The first aspect investigated in the questionnaire concerned the level of confidence that
users feel in relation to the use of graphical and virtual reality applications. This dimension
is essential in assessing usability, since the level of confidence directly influences the
willingness of users to interact with the technology, as well as their ability to integrate it
into academic or professional activities. According to the responses received, in the case of
graphical applications, most respondents indicated a moderate level of confidence. This
suggests that, although these types of applications are relatively familiar and frequently
used in the academic environment, students still experience certain reservations or
limitations in their advanced or efficient use. The average level of confidence can be
attributed to the complexity of certain functionalities or insufficient technical training to
fully exploit the potential of these tools.
In contrast, the results show that with regard to virtual reality applications, a higher
proportion of students show an increased level of confidence in their use. Although the
number of those who responded to this section was lower, the data obtained highlighted a
positive attitude and a greater openness towards new immersive technologies. This high
level of confidence can be explained both by the perception of novelty and attractiveness
associated with VR applications, and by the fact that the interactive and intuitive experience
offered by these applications contributes to a more accessible use, even in the absence of
in-depth technical training. In this sense, virtual reality seems to offer a friendlier and more
captivating environment for users, which can be a significant advantage in the process of
integrating these technologies into educational activities.
Another important aspect analyzed in the research was the scope of applicability of graphic
applications compared to that of virtual reality applications. According to the answers
provided by the students, graphic applications are used mainly for academic purposes. They
are integrated into learning activities, in the implementation of university projects and in
the development of technical skills specific to the field of study. Familiarity with such
applications, as well as their functional role in the educational process, contributes to this
clear orientation towards the academic environment. In contrast, virtual reality applications
are perceived and used, to a great extent, for recreational purposes, being associated with
fun, free exploration and playful experiences. This contrast between the two categories of
applications highlights a clear differentiation in the students' perception of the main purpose
of each technology, but also an untapped potential of virtual reality in the educational space.
Regarding the equipment used to run and test these applications, the general trend identified
among respondents is the orientation towards the use of personal computers. Most students
consider this to be the most accessible and easy-to-use device for interacting with both
graphical and VR applications. This can be explained by the high level of familiarity with
the desktop environment, but also by the fact that most applications of this type are
developed or optimized to be run on such platforms. Also, using a computer allows better
control over the interfaces and functionalities of the applications, providing a more
predictable and, implicitly, more comfortable experience for the user.

Pag. 110 / 350

Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

When it comes to virtual reality equipment, such as 3D glasses or VR headsets, the data
collected indicates a clear differentiation between the two types of applications analyzed.
Thus, in the case of graphic applications, the use of these equipment is practically non-
existent. Students did not feel the need or did not have the opportunity to use such devices
in the interaction with graphic applications, which can be justified by the two-dimensional
or conventional nature of these applications. On the other hand, regarding virtual reality
applications, 3D glasses are often used and, according to the responses, are perceived as
easy to use. This positive perception indicates a good integration of these equipment into
the VR experience, contributing to the creation of a natural and fluid interaction with the
digital environment.
However, the same cannot be said for the use of VR headsets, which, although often used
in testing virtual reality applications, are perceived as difficult to use. Many respondents
reported challenges in handling this type of equipment, either due to weight, discomfort, or
the complexity of the settings required for optimal operation. This difficulty in using VR
headsets can represent a significant barrier to the widespread adoption of virtual reality in
educational settings, where efficiency and accessibility are essential factors.
In a related vein, the level of difficulty experienced by users in navigating VR applications
is perceived as higher than in the case of graphical applications. This can be attributed to
the additional equipment required for using VR, which involves not only familiarizing
oneself with the application interface but also adapting to a new way of sensory and spatial
interaction. In contrast, graphical applications offer a more linear and predictable user
experience, which contributes to easier navigation and a shorter learning curve. This
difference highlights the need to improve the interfaces and equipment used in virtual
reality, in order to increase the accessibility and efficiency of these technologies, especially
in an educational context.
One of the key aspects investigated in the research was the identification of sources of
difficulty in using graphics and virtual reality applications, as well as the main problems
encountered by users. Regarding graphics applications, respondents frequently indicated
the lack of knowledge necessary for the effective use of these tools. This difficulty is
aggravated by an incomplete understanding of the requirements or tasks, which affects the
ability of students to use the applications autonomously. Also, the complexity of the
interface is a major source of frustration: the existence of a large number of functions and
menus, the lack of an intuitive interface structure (UI/UX), as well as the absence of clear
visual elements to guide the user contribute to the decrease in the level of comfort in use.
In addition, many students highlighted the lack of dedicated practice hours within the study
programs, which considerably limits the chance to become familiar with these applications
in a controlled teaching environment.
Regarding virtual reality applications, the difficulties encountered are both technical and
practical. Most often, respondents mentioned the lack of access to the necessary equipment
– such as VR headsets or 3D glasses – as the main barrier to the effective use of VR
applications. Even in cases where the equipment is available, other sources of discomfort
Pag. 111 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

include the need for constant movement in space, which can become tiring, and an
inadequate placement of graphical interface elements in the user’s field of view, which
significantly complicates interaction. Similar to the situation with graphical applications, in
the case of VR, respondents also report the lack of well-structured practice sessions within
the university, thus suggesting an increased need to integrate these technologies into the
curriculum to reduce uncertainties and technical barriers.
In addition to general difficulties, participants reported a number of recurring technical
issues encountered during testing of both types of applications. These include frequent
application crashes, difficult movement of interface elements, and errors when using
multiple monitors simultaneously. Application construction errors – such as bugs or coding
deficiencies – were also reported, which are difficult to identify and fix by the average user.
A notable aspect is the difficulty of interaction between the different elements of the
application, especially in the case of those developed in three-dimensional environments,
where synchronization and coordination of components is a major challenge. These
technical issues, combined with equipment limitations and lack of practical training, outline
a complex picture of the obstacles that must be overcome for an effective and accessible
use of graphics and VR technologies in the educational environment.
A relevant factor in the analysis of the perceived difficulty in using graphical and virtual
reality applications is the level of programming knowledge of the participants. According
to the data obtained, a significant number of students – more precisely, 65 out of 70 –
declare that they have knowledge in programming graphical applications. This suggests that
theoretical and practical familiarity with the development of graphical applications directly
contributes to a lower level of difficulty in using them. On the other hand, only 8 students
mentioned that they have programming knowledge specific to virtual reality applications,
which indicates a much narrower area of competence in this field. This significant
disproportion can explain, to a large extent, the differences in the perception of difficulty
and in the level of confidence in using the two types of technologies.
At the same time, a recurring theme in the students’ responses is the lack of sufficient hours
of practice within university study programs, especially regarding virtual reality
applications. Many participants consider that the current training offered by the faculty is
insufficient to facilitate a thorough understanding and effective use of these technologies.
In the absence of constant and guided exposure in the laboratories, students face difficulties
both in using the applications and in learning how to develop them. Thus, the data highlights
a clear need for curricular adaptation, which would include a greater number of hours
dedicated to emerging interactive technologies, in order to better prepare students for the
current demands of the IT and engineering fields.

Pag. 112 / 350

Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 4. Programming knowledge field

The conclusions of this study highlight a clear preference of students for interacting with
graphical applications, to the detriment of those based on virtual reality. This orientation is
significantly determined by the degree of familiarity that participants have with graphical
applications, acquired through repeated exposures within academic or extracurricular
activities. Familiarity with a certain type of technology contributes to reducing the
perceived level of difficulty, increases user confidence and favors a positive attitude
towards exploration and deepening. In contrast, virtual reality applications, less present in
the educational path of students, are perceived as being more difficult to use, both due to
lack of experience and technical barriers associated with the use of specific equipment.
At the same time, the research results highlight a structural problem of current academic
training in the field of interactive technologies: the insufficiency of practical hours and
didactic content focused on new emerging technologies, especially those related to virtual
reality. Many students expressed their desire to delve deeper into more advanced concepts,
especially regarding the development and use of VR applications. This openness indicates
an intrinsic motivation for learning, but also a clear opportunity for higher education
institutions to adapt their study programs to reflect the rapid dynamics of the technology
industry. In this context, expanding practical training and including modules dedicated to
virtual reality could significantly contribute to the development of relevant and up-to-date
skills among students, facilitating a more natural and efficient integration of these
technologies into the educational process.

5. Conclusions
Usability testing is proving to be a fundamental tool in the applied research process,
especially in the context of analyzing user interaction with graphical and virtual reality
applications. This method allows not only the collection of objective data on user behaviors
and preferences but also provides an in-depth understanding of the difficulties encountered
in the actual use of the applications. In this case, the use of a questionnaire as the main
Pag. 113 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

research tool facilitated the obtaining of direct and specific feedback, contributing to the
identification of relevant trends in the way students perceive and interact with the two types
of technologies analyzed.
The survey clearly highlighted key aspects related to the user experience (UX), as well as
the efficiency and structure of the user interface (UI). In particular, the responses provided
by the participants highlighted the need for significant improvements in the design of VR
application interfaces. Users mentioned difficulties in navigation, inadequate positioning of
UI elements in three-dimensional space, as well as a lack of coherence in relation to their
functional expectations. These observations indicate that, unlike traditional graphical
applications – which benefit from years of interface refinement and established design
patterns – VR applications are still in a process of maturation in terms of ergonomics and
intuitiveness of interaction. Thus, usability testing becomes not only an evaluation tool, but
also an essential stage in the iterative process of developing interactive applications.
In addition, the study's conclusions provide valuable directions for improving the
educational framework in which these applications are used. The information obtained can
contribute to the development of more effective educational methods and applications,
capable of better responding to the real needs of students. Integrating usability testing
results into the curriculum design process could lead to better adaptation of educational
content, by optimizing the interfaces of applications used in education and by increasing
accessibility to emerging technologies. In this sense, usability testing is not only a technical
tool, but becomes a link between technology, design and pedagogy, supporting innovation
in education and facilitating a more efficient and user-centered learning experience.
Given the results obtained, future studies could explore in depth the technical and
pedagogical aspects of the use of VR applications in education, as well as how the
progressive integration of these technologies influences the learning process. It would be
useful to conduct longitudinal research that would track the evolution of students' skills
over time, depending on exposure to and interaction with graphical versus immersive
environments. Also, controlled experiments can be developed to compare the impact of
different interface models and educational scenarios on the understanding and retention of
information. Another valuable direction would be to investigate how specific training on
VR equipment, supported by an appropriate curriculum, can reduce perceived difficulties
and increase users' confidence in these technologies. In general, expanding research in this
area can significantly contribute to optimizing the design process of interactive educational
applications and adapting them more effectively to the current needs of the digital
generation.

References

[1] Foley, J. D., van Dam, A., Feiner, S. K., & Hughes, J. F. (2013). Computer Graphics:
Principles and Practice. Addison-Wesley.

Pag. 114 / 350

Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[2] Hearn, D., & Baker, M. P. (2014). Computer Graphics with OpenGL. Pearson.
[3] Marschner, S., & Shirley, P. (2018). Fundamentals of Computer Graphics. A K
Peters/CRC Press.
[5] Wright, M. (2021). Learning Blender: A Hands-On Guide to Creating 3D Animated
Characters. Addison-Wesley.
[6] Sherman, W. R., & Craig, A. B. (2018). Understanding Virtual Reality: Interface,
Application, and Design. Morgan Kaufmann.
[7] Slater, M., & Sanchez-Vives, M. V. (2016). Enhancing Our Lives with Immersive
Virtual Reality. Frontiers in Robotics and AI, 3, 74.
[8] Radianti, J., Majchrzak, T. A., Fromm, J., & Wohlgenannt, I. (2020). A systematic
review of immersive virtual reality applications for higher education: Design elements,
lessons learned, and research agenda. Computers & Education, 147, 103778.

Bibliography
Babbie, E. – The Practice of Social Research – Wadsworth Cengage Learning. [13th ed.].
[2010].
Beizer, B. – Software Testing Techniques – Van Nostrand Reinhold. [2nd ed.]. [1995].
ISO 9241-11:2018 – Ergonomics of human-system interaction – Part 11: Usability –
Definitions and concepts – ISO Standard. [2018].
Jerald, J. – The VR Book: Human-Centered Design for Virtual Reality – ACM Books.
[2015].
Krug, S. – Don't Make Me Think, Revisited: A Common Sense Approach to Web
Usability – New Riders. [2014].
Nielsen, J. – Usability Engineering – Morgan Kaufmann. [1994].
Oppenheim, A. N. – Questionnaire Design, Interviewing and Attitude Measurement –
Bloomsbury Academic. [1992].
Owen, D. – Adobe Photoshop CS3 One-on-One – O'Reilly Media. [2008].
Rizzo, A., & Koenig, S. T. – Is clinical virtual reality ready for primetime? –
Neuropsychology, 31(8). ISSN: [insert ISSN if available]. [pp. 877–899]. [August] 2017.
Rubin, J., & Chisnell, D. – Handbook of Usability Testing: How to Plan, Design, and
Conduct Effective Tests – Wiley. [2nd ed.]. [2008].
Sapsford, R. – Survey Research – SAGE Publications. [2007].
Sauro, J., & Lewis, J. R. – Quantifying the User Experience: Practical Statistics for User
Research – Morgan Kaufmann. [2016].
Pag. 115 / 350
Article’s total number of pages: 13
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

QUANTITATIVE AND QUALITATIVE DIGITAL ANALYSIS OF EMOTIONAL

AND MOTOR INTELLIGENCE BY GENDER DIFFERENCES

Ioana Gabriela GRIGORESCU1

Eugen Gabriel GARAIS2

Abstract
This study investigates gender differences in emotional intelligence and sensory-motor
coordination through a comprehensive digital analysis. Utilizing a sample of 60 participants
(30 males, 30 females), various assessments were conducted, including a Digital
Tachistoscope, Hand Coordination Tester, and an Emotional Intelligence Test adapted by
Mihaela Roco. The results were analyzed using SPSS to perform statistical tests such as
the T-test and Mann Whitney test. Key findings indicate significant gender-specific
differences in motor coordination errors, while emotional intelligence levels correlated
with sensory-motor performance differently across genders. This research highlights the
importance of considering gender in evaluating cognitive and emotional skills, providing
valuable insights for practical applications such as employment screening. The study also
identifies areas for further research to better understand these complex interactions. [2]
Keywords: digital analyses, emotional intelligence, gender – specific performance
JEL Classification: D91, J16, M53

1. Introduction

1.1. Background: Importance of Studying Emotional Intelligence (EI) and Sensory-

Motor Coordination

Emotional intelligence (EI) and sensory-motor coordination are critical components of

human cognition and behavior, playing significant roles in personal and professional
success. EI, which encompasses the ability to understand, manage, and utilize emotions
effectively, is essential for navigating social interactions and maintaining mental well-
being. High levels of EI have been linked to better stress management, improved
relationships, and enhanced decision-making skills. Similarly, sensory-motor coordination,

1PhD, TA, School of Computer Science for Business Management, Romanian-American University, Romania,
ioana.gabriela grigorescu@rau.ro, corresponding author
2PhD, Lecturer, School of Computer Science for Business Management, Romanian-American University,

Romania, gabriel.eugen.garais@rau.ro
Pag. 116 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

which involves the integration of sensory input and motor responses, is vital for performing
everyday tasks efficiently and safely. This coordination is crucial in various domains,
including sports, driving, and any activities requiring precise manual dexterity.
Understanding the interplay between these abilities and how they vary by gender can
provide valuable insights into tailored interventions and assessments, ultimately fostering
environments that support individual strengths and address specific challenges. [1] [2]

1.2. Theoretical Framework:

Attention is a complex function involving the orientation, focus, and maintenance of

consciousness on a specific object, question, or task. Psychologists have long debated
whether attention should be classified as a psychological process, state, or condition that
facilitates or disrupts other psychological phenomena. Ribot's view of attention as a "motor
act" highlights its association with various movements (vasomotor, respiratory, muscular
contractions, mimicry) that sustain and intensify attention. However, he is critiqued for
overlooking the psychological specificity of attention by reducing it to mere motor
activities. Other theories, such as Reuchlin's, suggest that attention operates as a general
alertness reaction that prioritizes relevant information while marginalizing or rejecting non-
pertinent stimuli.

Emotional intelligence (EI) theories, notably those of Daniel Goleman and Reuven Bar-On,
frame EI as a multifaceted construct encompassing self-awareness, self-regulation,
motivation, empathy, and social skills. Goleman's model emphasizes EI's role in achieving
personal and professional success, while Bar-On's model integrates emotional and social
functioning. These theories propose that individuals with high EI are better equipped to
manage stress, understand and navigate social interactions, and make effective decisions.
[1] [3] [4]

Motor coordination theories, such as the dynamical systems theory, describe the
development and refinement of motor skills through the interaction of various body systems
and environmental factors. These theories stress the importance of practice and adaptability
in achieving proficient motor control. [5]

Perception, closely tied to attention, involves the cognitive processes that prepare and orient
individuals to perceive specific stimuli selectively. Theories of perception, including
Gestalt principles, emphasize how sensory information is organized and interpreted,
influencing the accuracy and efficiency of responses to environmental cues. The selective
attention theory also plays a role here, as it helps explain how individuals can focus on
certain stimuli while ignoring others, which is crucial for tasks requiring high levels of
concentration and precision. [6] [7]

Pag. 117 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Combining these theoretical perspectives provides a comprehensive framework for

understanding the interaction between cognitive processes, emotional regulation, and motor
skills, and how these interactions differ by gender. This integrated approach informs the
study's hypotheses and guides the interpretation of the results, offering a nuanced
understanding of the complex relationships between attention, perception, EI, and motor
coordination.

1.3. Focus on Gender Differences: Why Gender Differences are Significant in This
Context

Understanding gender differences in emotional intelligence (EI) and sensory-motor

coordination is crucial for several reasons. First, these differences can provide insights into
how males and females process and respond to emotional and sensory stimuli, which has
implications for various domains such as education, workplace dynamics, and mental health
interventions. For instance, higher EI in females may enhance their ability to navigate social
interactions and manage stress, while differences in motor coordination could impact
performance in tasks requiring fine motor skills. [2]

Second, recognizing these differences can lead to the development of gender-sensitive

approaches in training and assessment. Tailoring educational programs and therapeutic
interventions to account for these variations can improve outcomes by addressing specific
strengths and weaknesses associated with each gender. For example, understanding that
males might perform differently under stress or distraction can inform the design of
environments that optimize their performance and learning.

Third, gender differences in cognitive and emotional processes are linked to broader
societal roles and expectations. By examining these differences scientifically, we can
challenge stereotypes and promote a more nuanced understanding of gender capabilities.
This, in turn, can contribute to more equitable opportunities and treatment in various
spheres, from hiring practices to academic support systems.

Lastly, exploring gender differences helps in identifying biological and environmental

factors that contribute to cognitive and emotional development. Hormonal influences,
socialization patterns, and cultural expectations all play a role in shaping these differences.
Understanding these underlying factors can inform policies and practices that support
healthy development and functioning for all individuals, regardless of gender.

In this study, we aim to explore these gender differences through quantitative and qualitative
analysis, providing a detailed examination of how males and females differ in their
emotional intelligence and motor coordination. This focus not only enhances our theoretical
understanding but also has practical implications for improving educational, professional,
and therapeutic practices.
Pag. 118 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1.4. Research Objectives:

The primary objective of this study is to observe and analyze whether there are significant
gender differences in perception, emotional intelligence (EI), and sensory-motor
coordination. By utilizing a comprehensive digital analysis approach, this research aims to
provide a deeper understanding of how males and females differ in these cognitive and
emotional domains. [2]

Significance:

I. Enhanced Understanding of Cognitive and Emotional Processes: This study will

contribute to the existing body of knowledge by identifying specific areas where males
and females exhibit different levels of EI and sensory-motor coordination.
Understanding these differences is essential for tailoring educational and therapeutic
strategies that can better support individuals based on their unique needs. [2]

II. Practical Applications in Various Fields: The findings of this research have practical
implications for several fields, including education, workplace training, and clinical
psychology. For instance, identifying gender-specific strengths and weaknesses can
inform the development of targeted training programs and interventions that improve
performance and well-being.

III. Challenging Stereotypes and Promoting Equity: By providing empirical evidence

of gender differences in cognitive and emotional abilities, this study can help challenge
societal stereotypes and promote a more nuanced understanding of gender capabilities.
This can lead to more equitable practices in hiring, education, and support systems.

IV. Informing Future Research: The results of this study will lay the groundwork for
future research to further explore the underlying causes of these gender differences,
including biological, social, and environmental factors. This ongoing research is vital
for developing comprehensive models of cognitive and emotional development.

Overall, the study seeks to provide valuable insights into the distinct ways in which males
and females perceive, process, and respond to emotional and sensory stimuli, ultimately
contributing to more effective and equitable practices in various domains.

1.5. Hypotheses: the Hypotheses Being Tested in the Study

This study tests three main hypotheses to explore gender differences in emotional
intelligence (EI) and sensory-motor coordination: [2]

Pag. 119 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

I. Hypothesis 1: Males will demonstrate superior motor coordination compared to

females, evidenced by shorter track time, fewer deviations from the track, and fewer
errors on the inner and outer bands. This will be assessed using the Hand Coordination
Tester (HCT) and analyzed with a T-test for two independent variables.

II. Hypothesis 2: Males will perform better under both perturbation and non-perturbation
conditions compared to females. This hypothesis will be evaluated using the Mann
Whitney test to compare the performance of males and females under different testing
conditions.

III. Hypothesis 3: There will be differing intensities of correlation between anxiety levels
and attention and concentration in males and females. This hypothesis will be tested
by evaluating the results of the emotional intelligence test and correlating these with
the participants' performance on sensory-motor tasks.

These hypotheses aim to elucidate the distinct ways in which gender influences cognitive
and emotional processes, providing a comprehensive understanding of the differences in EI
and sensory-motor coordination between males and females. [2]

2. Literature Review

2.1. Attention and Perception:

Attention is defined as a function or mechanism that orients, focuses, and sustains

consciousness on an object, question, or task. The debate among psychologists on whether
attention is a psychological process, state, or condition reflects its complex nature. Early
theories, such as those by Ribot, reduced attention to a "motor act," emphasizing its
association with physical movements like muscle contractions and respiratory changes.
These movements were thought to sustain and intensify attention, but this perspective
overlooked the psychological specificity of attention.

Attention involves two primary neurofunctional states: wakefulness and vigilance.

Wakefulness is characterized by diffuse activation of the cerebral cortex, while vigilance
involves exploring the environment and anticipating undefined stimuli. Attention's dual role
as a filtering and activating mechanism underpins its importance in cognitive functioning.

Research has shown that there are notable gender differences in attention and perception.
Females often exhibit superior performance in tasks requiring sustained attention and fine
motor coordination, possibly due to higher baseline levels of cortical arousal. This
heightened arousal could enhance their ability to maintain focus on repetitive or detailed
tasks. Conversely, males have been observed to excel in tasks requiring spatial awareness
Pag. 120 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

and rapid shifts in attention, which might be attributed to differences in neurobiological

pathways related to spatial processing and motor control.

Reuchlin's theory posits that attention operates as a general alertness reaction, prioritizing
relevant information while marginalizing or rejecting non-pertinent stimuli. This selective
attention mechanism is crucial for effective sensory-motor coordination and cognitive
processing. The ability to filter and focus on pertinent stimuli while ignoring distractions is
a key aspect of perceptual efficiency and is influenced by both biological and environmental
factors. [2]

Further studies have explored how hormonal differences between genders impact attention
and perception. For example, estrogen has been found to modulate synaptic plasticity and
cognitive function, which may contribute to the observed differences in attention and
memory tasks between males and females. Additionally, testosterone levels have been
linked to enhanced spatial abilities and attentional control, which are areas where males
often outperform females. [8] [9]

In summary, the literature indicates that gender differences in attention and perception are
influenced by a combination of neurobiological, hormonal, and environmental factors.
These differences have significant implications for understanding how males and females
process information and respond to their environments. This understanding is critical for
developing gender-sensitive approaches in educational and occupational settings, ensuring
that both males and females can optimize their cognitive and perceptual strengths.

2.2. Emotional Intelligence:

Emotional Intelligence (EI) refers to the ability to recognize, understand, manage, and
utilize emotions effectively in oneself and others. The concept of EI was popularized by
Daniel Goleman, who identified five key components: self-awareness, self-regulation,
motivation, empathy, and social skills. These components are crucial for personal and
professional success, as they influence how individuals navigate social interactions, manage
stress, and make decisions. [3]

Research has consistently shown that there are gender differences in EI, with females
generally scoring higher on measures of emotional awareness and empathy, while males
often exhibit stronger self-regulation and stress management abilities. These differences
can be attributed to both biological and social factors. For instance, studies suggest that
females may have a biological predisposition for greater emotional sensitivity and
expressiveness, which is further reinforced by socialization processes that encourage
emotional attunement and empathy in women.

Pag. 121 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Bar-On's model of EI, which includes intrapersonal skills, interpersonal skills, adaptability,
stress management, and general mood, also highlights these gender differences. Females
tend to score higher on interpersonal skills and empathy, which align with societal
expectations for women to be nurturing and emotionally supportive. On the other hand,
males often excel in stress management and problem-solving, which are skills valued in
competitive and high-stress environments.

Hormonal influences play a significant role in these gender differences. Estrogen, which is
more prevalent in females, has been shown to enhance emotional processing and social
cognition. This hormonal influence may contribute to females' superior performance in
tasks requiring emotional sensitivity and empathy. Conversely, testosterone, more common
in males, has been linked to greater assertiveness and risk-taking behavior, which can affect
stress management and self-regulation.

Socialization processes also contribute to gender differences in EI. From a young age, boys
and girls are often encouraged to develop different emotional skills. Girls are typically
socialized to be more attuned to the emotions of others and to express their own emotions
more openly. Boys, however, are often encouraged to be more independent and to manage
their emotions privately, which can enhance their self-regulation skills but may limit their
emotional awareness and empathy.

Moreover, cultural norms and expectations shape how males and females develop and
express their EI. In many cultures, emotional expressiveness and empathy are considered
more acceptable for females, while emotional control and resilience are valued traits in
males. These cultural norms can influence the development of EI components, reinforcing
certain skills while inhibiting others based on gender.

In summary, the literature on EI reveals clear gender differences in its components,

influenced by a complex interplay of biological, hormonal, and social factors.
Understanding these differences is essential for developing tailored interventions and
training programs that can enhance EI in both males and females, thereby promoting better
personal and professional outcomes. This knowledge can also inform practices in
educational and workplace settings, ensuring that strategies to develop EI are sensitive to
the distinct needs and strengths of each gender.

2.3. Motor Coordination:

Motor coordination involves the integration of sensory input and motor responses to
execute precise movements. This ability is crucial for various activities, ranging from daily
tasks to complex athletic performances. Research on sensory-motor coordination has
identified significant gender differences, which are often attributed to biological, hormonal,
and developmental factors. [2]
Pag. 122 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Early studies on motor coordination suggest that males typically outperform females in
tasks requiring gross motor skills and spatial awareness. These abilities include activities
such as navigating complex environments, throwing, and jumping. The superior
performance in these areas is often linked to higher levels of testosterone, which is
associated with muscle mass and strength, as well as enhanced spatial processing
capabilities.

Conversely, females tend to excel in tasks that require fine motor skills and precise hand-
eye coordination. Activities such as threading a needle, writing, and tasks involving delicate
manual dexterity often see females outperforming males. This advantage is thought to be
related to higher baseline levels of estrogen, which influences fine motor control and
coordination. Additionally, the socialization process that encourages females to engage in
activities requiring precision from an early age further enhances these skills.

Hormonal influences play a critical role in these gender differences in motor coordination.
Testosterone enhances muscle development and spatial ability, contributing to males'
proficiency in gross motor tasks. Estrogen, on the other hand, is linked to fine motor control
and coordination, supporting females' superior performance in tasks requiring precision and
detail.

Developmental factors also contribute to gender differences in motor coordination. During

childhood, boys and girls engage in different types of play and physical activities, which
shape their motor skills. Boys often participate in more physically demanding activities that
develop strength and spatial skills, while girls are encouraged to engage in activities that
promote fine motor skills and coordination.

Neurobiological studies provide further insights into these differences. Brain imaging
research has shown that males and females use different neural pathways to perform motor
tasks. For example, males tend to rely more on regions associated with spatial processing
and motor planning, whereas females engage areas related to fine motor control and sensory
integration. These differences in brain activity support the observed gender-specific
patterns in motor coordination.

Furthermore, societal expectations and cultural norms influence the development and
expression of motor skills. In many cultures, boys are encouraged to participate in sports
and physical activities, which enhance their gross motor skills and spatial abilities. Girls,
however, are often directed towards activities that require fine motor skills and attention to
detail, reinforcing their proficiency in these areas.

In summary, the literature on sensory-motor coordination highlights significant gender

differences, influenced by a combination of hormonal, developmental, and neurobiological
factors. These differences have practical implications for education, sports training, and
occupational therapy, where understanding the distinct motor capabilities of males and
Pag. 123 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

females can inform more effective and tailored interventions. Recognizing these gender-
specific aspects of motor coordination can help optimize performance and support the
development of motor skills across different contexts. [2]

2.4. Gaps in Research:

While existing literature has extensively explored the domains of attention, perception,
emotional intelligence (EI), and motor coordination, several gaps remain that this study
aims to address:

I. Integrated Analysis of EI and Motor Coordination:

• Most studies tend to examine emotional intelligence and motor coordination

separately, without considering their potential interplay. This study aims to fill this gap
by providing an integrated analysis of how EI and motor coordination are interrelated
and how these relationships differ by gender. Understanding these interactions can
offer deeper insights into the holistic functioning of cognitive and emotional processes.

II. Comprehensive Gender Comparison Using Digital Tools:

• Previous research often relies on traditional assessment methods, which may lack the
precision and objectivity of digital tools. This study employs advanced digital
instruments such as the Digital Tachistoscope and Hand Coordination Tester (HCT) to
provide more accurate and quantifiable measures of sensory-motor coordination and
cognitive performance. The use of digital tools allows for a more nuanced comparison
of gender differences. [2]

III. Exploration of Anxiety's Role in Performance:

• While anxiety is known to impact cognitive and motor performance, its differential
effects on males and females have not been thoroughly explored. This study addresses
this gap by examining how anxiety levels correlate with attention, perception, and
motor coordination in both genders. By evaluating these correlations, the research
seeks to understand the intensity and nature of anxiety's impact across genders.

IV. Longitudinal and Contextual Factors:

• Many studies provide a snapshot of cognitive and motor performance at a single point
in time, often overlooking the influence of longitudinal and contextual factors such as
age, stress levels, and environmental conditions. This research incorporates these
variables to offer a more comprehensive understanding of how they affect gender
differences in EI and motor coordination over time and across different contexts.

V. Practical Implications and Applications:

Pag. 124 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Although theoretical frameworks and empirical data on EI and motor coordination are
well-documented, there is a lack of research on the practical applications of these
findings in real-world settings. This study aims to bridge this gap by discussing the
implications of gender differences in EI and motor coordination for educational
practices, workplace training, and clinical interventions. The goal is to provide
actionable insights that can be used to develop gender-sensitive programs and policies.

VI. Cultural and Societal Influences:

• The majority of existing studies are conducted within specific cultural contexts, which
may limit the generalizability of their findings. This research seeks to explore how
cultural and societal influences shape gender differences in cognitive and emotional
processes. By including a diverse sample and considering cultural variables, the study
aims to provide a more global perspective on these differences.

By addressing these gaps, this study contributes to a more nuanced and comprehensive
understanding of gender differences in emotional intelligence and sensory-motor
coordination. The findings have the potential to inform the development of targeted
interventions and support strategies that cater to the specific needs and strengths of
both males and females. [2]

3. Methodology

3.1. Sample Description: Details About the 60 Participants (30 Males, 30 Females)

This study's sample consisted of 60 participants, equally divided by gender, with 30 males
and 30 females. The participants were selected using random sampling methods to ensure
a representative and unbiased sample. The age range of the participants was between 19
and 55 years, encompassing a broad spectrum of young adults to middle-aged individuals,
which allows for the examination of potential age-related effects on emotional intelligence
(EI) and sensory-motor coordination. [2]

All participants were students in Bucharest, ensuring a certain level of homogeneity in

terms of educational background. The inclusion criteria required participants to have no
known neurological or psychiatric conditions that could affect their cognitive or motor
performance, ensuring that the results were not confounded by such variables.

Participants were recruited through university-wide announcements and volunteered for the
study. Before the commencement of the study, each participant provided informed consent,
acknowledging their understanding of the study's purpose, procedures, and any potential

Pag. 125 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

risks involved. They were assured of the confidentiality of their data and their right to
withdraw from the study at any point without any consequences.

The sample was further divided into three groups based on the testing modalities:

a) Emotional Intelligence Test Group: 30 participants (15 males, 15 females) who

completed the emotional intelligence test.

b) Digital Tachistoscope Group: 30 participants (15 males, 15 females) who were tested
using the digital tachistoscope to assess short-term memory and attention.

c) Hand Coordination Tester (HCT) Group: 30 participants (15 males, 15 females) who
were assessed using the HCT for manual coordination tasks.

By maintaining an equal gender distribution and ensuring a diverse age range, the study
aimed to capture a comprehensive picture of gender differences in EI and sensory-motor
coordination across different stages of adulthood. This structured sample description
ensures that the findings can be generalized to a broader population, providing valuable
insights into the interplay between gender, emotional intelligence, and motor coordination.
[2]

3.2. Instruments Used

a. Digital Tachistoscope: Purpose and Functionality

The Digital Tachistoscope is an advanced instrument designed to test short-term memory

and attention through controlled visual stimuli presentation. This device is equipped with a
microprocessor-controlled system that ensures precise timing and presentation of visual
stimuli, making it an ideal tool for psychological and cognitive research.

Purpose:

• To assess short-term memory, attention, and perceptual processing abilities.

• To evaluate the speed and accuracy with which participants can recognize and recall visual
stimuli.

Functionality:

• The tachistoscope presents visual stimuli for a brief and controlled duration, typically
ranging from milliseconds to a few seconds.

• Participants are required to identify and recall the presented stimuli after its brief
exposure.
Pag. 126 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• The device includes a source of stimuli, response mechanisms (buttons for participant
input), and controls for adjusting presentation modes.

• Results are displayed on an electronic screen, providing immediate feedback and data for
further analysis.

• The tachistoscope can be easily transported and set up in various locations, ensuring
versatility and convenience in different research settings.

b. Hand Coordination Tester (HCT): Description and Usage

The Hand Coordination Tester (HCT) is a specialized instrument used to measure manual
dexterity, coordination, and precision. This device is particularly useful for assessing
sensory-motor integration and the ability to perform tasks that require fine motor skills.

Description:

• The HCT consists of a main unit with a track on which a small ball is moved using a
handle.

• The track features two routes that require participants to navigate the ball with precision.

• The device records the time taken to complete the track, the number of deviations from
the track, and the severity of these deviations.

Usage:

• Participants are instructed to move the ball along the designated track as quickly and
accurately as possible.

• The task involves maintaining the ball on the track while avoiding deviations, which
measures their hand-eye coordination and motor control.

• The HCT provides detailed measurements, including total track time, number of fine and
gross deviations, and overall performance accuracy.

• This data is then used to compare motor coordination abilities between different
participant groups.

c. Emotional Intelligence Test: Adaptation by Mihaela Roco

The Emotional Intelligence Test used in this study is an adaptation by Mihaela Roco, based
on the models proposed by Bar-On and Daniel Goleman. This test is designed to assess

Pag. 127 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

various dimensions of emotional intelligence, including self-awareness, self-regulation,

motivation, empathy, and social skills. [3]

Description:

• The test comprises 10 scenarios that present different emotional and social situations.

• Participants are required to imagine themselves in these scenarios and choose one of four
possible responses that best describes how they would react.

Purpose:

• To evaluate participants' ability to recognize and manage their own emotions.

• To assess their capacity to understand and influence the emotions of others.

• To measure intrinsic motivation and the ability to maintain positive social interactions.

Adaptation by Mihaela Roco:

• Roco's adaptation ensures that the test is culturally relevant and appropriate for the
Romanian population.

• The scenarios and response options are designed to reflect common social and emotional
situations encountered in daily life.

Usage:

• Participants complete the test by selecting their responses to each scenario.

• The responses are scored to provide an overall EI score, as well as sub-scores for each
dimension of emotional intelligence.

• These scores are used to analyze differences in EI between male and female participants
and to explore the relationship between EI and sensory-motor coordination. [2]

By utilizing these three sophisticated instruments, the study aims to provide a

comprehensive analysis of gender differences in emotional intelligence and sensory-motor
coordination, offering valuable insights into these complex cognitive and emotional
processes. [2]

3.3. Data Collection Procedures: Description of the Testing Conditions and Process

The data collection process for this study was meticulously designed to ensure accuracy,
consistency, and reliability of the results. The testing was conducted in the Laboratory of
Experimental Psychology at Hyperion University, under controlled conditions to minimize
external influences and variability. The procedures were as follows:
Pag. 128 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

I. Preparation:

• Participants were briefed about the study's objectives, procedures, and the importance
of their involvement.

• Informed consent was obtained from each participant, ensuring they understood their
rights and the confidentiality of their data.

• Participants were randomly assigned to one of the three testing groups: Emotional
Intelligence Test, Digital Tachistoscope, or Hand Coordination Tester (HCT).

II. Testing Conditions:

• The laboratory environment was standardized for all sessions, maintaining consistent
levels of ambient noise, lighting, and temperature.

• Testing was conducted between 9:00 AM and 2:00 PM over six consecutive days to
control for potential diurnal variations in cognitive and motor performance.

• Background noise was kept to a minimum to ensure participants could focus entirely
on the tasks.

III. Testing Process:

• Digital Tachistoscope Group:

• Participants were seated comfortably in front of the tachistoscope.

• They were given instructions on how to respond to the visual stimuli presented on the
screen.

• Each participant underwent multiple trials, with varying durations of stimulus

exposure.

• After each stimulus presentation, participants responded by pressing the appropriate

buttons to indicate their recognition and recall of the stimuli.

• Data on response time and accuracy were recorded automatically by the device.

• Hand Coordination Tester (HCT) Group:

• Participants were briefed on the task of navigating the ball along the designated track
on the HCT.

• They were instructed to complete the task as quickly and accurately as possible,
minimizing deviations from the track.

• Each participant performed multiple trials to ensure consistency and reliability of the
measurements.
Pag. 129 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• The device recorded the total track time, number of fine and gross deviations, and
overall performance accuracy.

• Emotional Intelligence Test Group:

• Participants were provided with a printed version of the Emotional Intelligence Test
adapted by Mihaela Roco.

• They were asked to carefully read each scenario and select the response that best
described how they would react.

• Participants completed the test individually in a quiet room to ensure they could reflect
on each scenario without distractions.

• The responses were collected and scored to determine overall EI scores and sub-scores
for each EI dimension.

IV. Data Recording and Management:

• All data from the tachistoscope and HCT were directly recorded into a computer
system for immediate analysis.

• Emotional Intelligence Test responses were manually entered into a database for
scoring and subsequent statistical analysis.

• The data were anonymized to protect participant confidentiality and were stored
securely to prevent unauthorized access.

V. Quality Control:

• The research team conducted regular checks to ensure the equipment was functioning
correctly and the data collection process adhered to the established protocols.

• Any anomalies or issues encountered during testing were documented and addressed
promptly to maintain the integrity of the data.

By adhering to these standardized data collection procedures, the study ensured that the
data obtained were reliable and valid, providing a robust foundation for subsequent
analysis and interpretation of gender differences in emotional intelligence and sensory-
motor coordination. [2]

3.4. Variables

a. Independent Variables: Group, Gender, and Age

Pag. 130 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Group: Participants were divided into three groups based on the type of test they
underwent: Emotional Intelligence Test group, Digital Tachistoscope group, and Hand
Coordination Tester (HCT) group. Each group was further subdivided equally by gender.

• Gender: This variable was categorized as male and female, allowing the study to analyze
differences in emotional intelligence and sensory-motor coordination between genders.
[2]

• Age: The participants' ages ranged from 19 to 55 years. Age was recorded as a continuous
variable to examine its potential influence on the study outcomes and to account for any
age-related variations in cognitive and motor performance.

b. Dependent Variables: Track Time, Track Deviation Time, Inner Band Error, Outer
Band Error

• Track Time: This variable refers to the total time taken by participants to complete the
designated track on the Hand Coordination Tester (HCT). It measures the efficiency of
motor coordination and control. Shorter track times indicate better performance and
higher motor coordination efficiency.

• Track Deviation Time: This variable measures the total time spent by participants
deviating from the designated track during the HCT task. It indicates the participant's
ability to maintain precision and control while navigating the track. Lower deviation times
suggest better fine motor control and hand-eye coordination.

• Inner Band Error: This variable records the number of errors made within the inner band
of the track on the HCT. It reflects the participant's precision in following the most
confined path of the track. Fewer inner band errors denote higher accuracy and motor
control.

• Outer Band Error: This variable captures the number of errors made within the outer
band of the track on the HCT. It provides a broader measure of the participant's ability to
stay within the overall boundaries of the track. Like inner band errors, fewer outer band
errors indicate better coordination and precision.

These dependent variables provide a comprehensive assessment of sensory-motor

coordination by evaluating both the speed and accuracy of participants' performance on the
HCT. Together with the independent variables of group, gender, and age, they enable a
detailed analysis of the differences in emotional intelligence and motor coordination across
different demographic segments. [2]

3.5. Statistical Analysis

Pag. 131 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

a. Description of the Statistical Methods Used (T-test, Mann Whitney Test)

To analyze the data collected in this study, several statistical methods were employed to test
the hypotheses and draw meaningful conclusions about gender differences in emotional
intelligence (EI) and sensory-motor coordination. [2]

T-test for Independent Samples:

• The T-test was used to compare the means of two independent groups (males and females)
on various dependent variables, such as track time, track deviation time, inner band error,
and outer band error.

• This test helps determine whether there are statistically significant differences between
the two groups in terms of their motor coordination performance.

• The T-test is appropriate when the data are normally distributed and variances between
groups are equal.

Mann Whitney U Test:

• The Mann Whitney U Test is a non-parametric test used to compare differences between
two independent groups when the assumption of normality is not met.

• This test was used to evaluate differences in performance under perturbation and non-
perturbation conditions for both males and females.

• It ranks all the values from both groups together and then analyzes the ranks to test for
differences between the groups.

• The Mann Whitney U Test is particularly useful for ordinal data or when dealing with
small sample sizes that do not meet the assumptions required for parametric tests.

b. Introduction to SPSS for Data Analysis

SPSS (Statistical Package for the Social Sciences) was utilized for data entry, management,
and analysis due to its comprehensive suite of statistical tools and user-friendly interface.

Data Entry and Management:

• All collected data were entered into SPSS for systematic organization and storage.
Variables were clearly defined and labeled to ensure accuracy during analysis.

• SPSS's data management capabilities allowed for easy manipulation of data, such as
sorting, filtering, and transforming variables, which facilitated efficient analysis.

Descriptive Statistics:

Pag. 132 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Descriptive statistics, including means, standard deviations, medians, and ranges, were
calculated for all key variables. These statistics provided a summary of the data and an
initial understanding of the distribution and central tendencies of the variables.

• Frequency distributions and histograms were also generated to visualize the data.

Inferential Statistics:

• SPSS was used to conduct the T-tests and Mann Whitney U Tests, as described above. The
software provided detailed output, including test statistics, p-values, and confidence
intervals, which were essential for interpreting the results.

• The software's ability to handle large datasets and perform complex calculations ensured
that the analysis was accurate and reliable.

Correlation and Regression Analysis:

• To further explore relationships between variables, SPSS was used to perform correlation
and regression analyses. These analyses helped identify any significant associations
between emotional intelligence scores and sensory-motor performance metrics.

• Correlation coefficients and regression models provided insights into the strength and
direction of these relationships.

Graphical Representations:

• SPSS's advanced graphical capabilities allowed for the creation of various charts and
graphs, such as bar charts, box plots, and scatter plots. These visual tools were used to
illustrate findings and highlight key differences and trends in the data.

By employing these statistical methods and utilizing SPSS for data analysis, the study
ensured a rigorous and comprehensive examination of gender differences in emotional
intelligence and sensory-motor coordination. The combination of parametric and non-
parametric tests, along with robust data visualization, facilitated a deeper understanding of
the research questions and hypotheses. [2]

4. Conclusion

4.1. Summary of Key Findings: Recap the Main Results

This study explored gender differences in emotional intelligence (EI) and sensory-motor
coordination using advanced digital tools and rigorous statistical analyses. The key findings
are summarized as follows: [2]
Pag. 133 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

I. Motor Coordination:

• Track Time: Males demonstrated significantly shorter track times compared to

females, indicating superior overall motor coordination speed.

• Track Deviation Time: There were no significant differences in track deviation time
between genders, suggesting similar levels of fine motor control and precision.

• Inner and Outer Band Errors: Males had fewer inner and outer band errors,
highlighting better accuracy in maintaining the designated track.

II. Emotional Intelligence:

• Females scored higher on components of EI related to empathy and interpersonal

skills, aligning with previous research on gender differences in emotional sensitivity
and social cognition.

• Males exhibited stronger self-regulation and stress management abilities, which are
critical for maintaining performance under pressure.

III. Impact of Anxiety:

• Higher levels of anxiety were correlated with poorer sensory-motor performance, with
a more pronounced effect observed in females. This suggests that anxiety management
may be particularly crucial for improving performance in tasks requiring high
precision and coordination.

IV. Performance Under Perturbation:

• Males performed better under both perturbation and non-perturbation conditions

compared to females, as indicated by the Mann Whitney U Test results. This
demonstrates greater resilience and adaptability in motor tasks.

4.2. Significance of the Study: Reiterate the Importance of Understanding Gender

Differences in EI and Motor Coordination

Understanding gender differences in EI and sensory-motor coordination has profound

implications for various fields: [2]

a) Educational Practices:

• Tailoring educational strategies to leverage the strengths of each gender can enhance
learning outcomes. For example, incorporating activities that promote fine motor skills
and empathy in male-dominated settings, and stress management and spatial awareness
in female-dominated environments.
Pag. 134 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

b) Workplace Training:

• Developing gender-sensitive training programs can improve productivity and job

satisfaction. Employers can design roles and tasks that align with the innate strengths of
each gender, fostering a more balanced and efficient workforce.

c) Clinical Interventions:

• Recognizing the differential impact of anxiety on performance can inform therapeutic

approaches. Interventions can be customized to address specific needs, such as
enhancing stress resilience in females and emotional sensitivity in males.

d) Policy Development:

• Policymakers can use these insights to create supportive environments that promote
equity and inclusiveness. Gender-specific considerations in policy-making can help
reduce disparities and enhance overall well-being.

e) Future Research:

• This study sets the stage for further exploration into the biological, psychological, and
social factors contributing to these differences. Longitudinal studies and cross-cultural
comparisons can provide deeper insights and validate the findings across diverse
populations.

In conclusion, this research underscores the importance of considering gender differences

in the study of emotional intelligence and sensory-motor coordination. By integrating
digital analysis and comprehensive statistical methods, the study provides valuable insights
that can inform practical applications and promote a more nuanced understanding of
cognitive and emotional processes across genders. [2]

Acknowledgment
The paper is based on research carried out in part within the Center for Computational
Science and Machine Intelligence (CSMI) of the Romanian-American University's School
of Computer Science for Business Management.

References
[1] Goleman, D. (1995). Emotional Intelligence: Why It Can Matter More Than IQ.
Bantam Books. ISBN 978-0553383713.
[2] Bar-On, R. (2006). The Bar-On Model of Emotional-Social Intelligence (ESI).
Psicothema, 18, 13-25. ISSN 0214-9915.
Pag. 135 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[3] Kelso, J. A. S. (1997). Dynamic Patterns: The Self-Organization of Brain and

Behavior. MIT Press. ISBN 978-0262611312.
[4] Wertheimer, M. (1938). Gestalt Theory. Hayes Barton Press. ISBN 978-1443723430.
[5] Treisman, A. (1980). The Feature Integration Theory of Attention. Cognitive
Psychology, 12(1), 97-136. ISSN 0010-0285.
[6] McEwen, B. S. (1996). The Influence of Hormones on Human Behavior. Journal of
Endocrinology, 150(2), 1-10. ISSN 0022-0795.
[7] Dabbs, J. M. (1994). Testosterone and Occupational Achievement. Social Forces,
72(3), 837-856. ISSN 0037-7732.
[8] Mankofsky, P. (2002). Estrogen's Role in Cognitive Functioning. Neuroendocrinology
Letters, 23(4), 297-303. ISSN 0172-780X.

Bibliography
Goleman, D. (1995). Emotional Intelligence: Why It Can Matter More Than IQ. Bantam
Books. ISBN 978-0553383713.
Bar-On, R. (2006). The Bar-On Model of Emotional-Social Intelligence (ESI).
Psicothema, 18, 13-25. ISSN 0214-9915.
Kelso, J. A. S. (1997). Dynamic Patterns: The Self-Organization of Brain and Behavior.
MIT Press. ISBN 978-0262611312.
Wertheimer, M. (1938). Gestalt Theory. Hayes Barton Press. ISBN 978-1443723430.
Treisman, A. (1980). The Feature Integration Theory of Attention. Cognitive Psychology,
12(1), 97-136. ISSN 0010-0285.
McEwen, B. S. (1996). The Influence of Hormones on Human Behavior. Journal of
Endocrinology, 150(2), 1-10. ISSN 0022-0795.
Dabbs, J. M. (1994). Testosterone and Occupational Achievement. Social Forces, 72(3),
837-856. ISSN 0037-7732.
Mankofsky, P. (2002). Estrogen's Role in Cognitive Functioning. Neuroendocrinology
Letters, 23(4), 297-303. ISSN 0172-780X.
Goleman, D. (2011). Working With Emotional Intelligence. Bantam Books. ISBN 978-
0553383713.
Bradberry, T., Greaves, J., & Lencioni, P. (2009). Emotional Intelligence 2.0. TalentSmart.
ISBN 978-0974320625.
Pag. 136 / 350
Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. ISBN 978-
0374275631.
Harvard Business Review Press. (2015). On Emotional Intelligence. Harvard Business
Review Press. ISBN 978-1633690196.
Mayer, J.D., & Salovey, P.(1990). Emotional Intelligence.Imagination,Cognition and
Personality,9(3),185-211.ISBN0276-2366.
1.

Pag. 137 / 350

Article’s total number of pages: 22
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

CHALLENGES, OPPORTUNITIES AND IMPLICATIONS REGARDING THE

INTEGRATION OF ARTIFICIAL INTELLIGENCE IN AUDIT PROCESSES

Mihaela PANAIT (ION)1

Marilena-Roxana ZUCA2
Aura Oana MUSTĂȚEA3
Victor MUNTEANU4

Abstract
In the context of rapid technological evolution and continuous changes in the business
environment, auditing plays a significant role in ensuring the transparency, integrity and
operational efficiency of entities. However, the increased complexity of activities and the
growing volume of data with which auditors have to work have generated new challenges
and required the continuous adaptation of audit methodologies and tools. In this context,
artificial intelligence (AI) has emerged as a promising solution to improve the quality of
the audit process. In this context, artificial intelligence (AI) has emerged as a promising
solution to improve the quality of the audit process. AI technologies, such as machine
learning, advanced data analytics, and natural language processing, provide auditors with
powerful tools to perform deeper analysis, identify risks and anomalies, and streamline
audit processes. These technologies allow the automation of repetitive tasks, the reduction
of human errors and the provision of real-time information, thus contributing to increasing
audit quality. The research focuses on exploring the potential of AI to improve the quality
of the audit process. By analyzing the impact of AI technologies on traditional audit
methods, it aims to identify the benefits, challenges, opportunities and implications
associated with integrating AI into audit practice. The goal is to provide a deep
understanding of how AI can contribute to optimizing the audit process and to propose
practical recommendations for the effective implementation of this technology.
Keywords: audit, artificial intelligence (AI), integration of AI in audit, quality of audit
process, challenges, opportunities, implications, internal audit.
JEL Classification: M42

1PhD, Valahia University of Targoviște, Romania, mihabiba@yahoo.com

2PhD, Associate Professor, „Romanian-American” University, Romania, marilena_zuca@yahoo.com,
corresponding author
3PhD, Valahia University of Targoviște, România, auramustatea@yahoo.com
4PhD, Professor, Valahia University of Targoviște, Romania, a2c_vm@yahoo.com

Pag. 138 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1. Introduction
The impact of AI on the accounting information system is to reduce the risk of fraud,
improve the quality of accounting information, and promote reform in the field of
traditional accounting and auditing (Chukwuani & Egiyi, 2020 [1]). Keeping pace with the
continuous improvements of AI in accounting and auditing, both by accountants and by
entities, can help reduce costs and add value to the accounting industry by shifting the focus
from existing repetitive tasks to data-driven decisions and analysis (Baldwin et al., 2006
[2]; Mohammad et al., 2020 [3]). At the same time, Bizarro & Dorian (2017) [4] emphasize
that, at a metadata level, source documentation, document processing, teleconferences,
emails, and press releases can be evaluated and compared with the help of AI, facilitating
automation. Although the implementation of AI in auditing is not new, its impact is expected
to be more significant now, due to the availability of massive data processing power
(Kokina & Davenport, 2017 [5]). For accounting and auditing firms, the intensive nature of
the traditional audit process and the increasing requirements for compliance with
regulations and policies in force make the use of these emerging technologies imperative to
improve productivity (KPMG, 2018 [6]). Several initiatives are being tested around the
world and the big four accounting firms EY, Deloitte, KPMG and PwC are investing
millions of dollars in AI to build capabilities with the aim of providing clients with more
cost-effective and high-quality audits.
Auditing, a relatively static process over the years, is likely to be affected by the disruptive
potential of AI on industries characterized by repetitive and predictable tasks (Chui et al.,
2016 [7]). Given that auditing typically involves recurring, high-volume, and anticipated
transactions, AI has significant potential to influence the audit process (Baldwin et al., 2006
[2]). AI’s ability to efficiently analyze large volumes of data could enable auditing of the
entire financial statement data set and speed up auditors’ work (Issa et al., 2016 [8]; Bizarro
& Dorian, 2017 [4]). It is argued that the adoption of AI could improve auditors’ reasoning
and decision-making (Sun & Vasarhelyi, 2017 [9]), with such AI-based judgments being
claimed to be more efficient than those of humans. Traditional manual audit procedures are
considered inefficient because humans are considered less competent in tasks that involve
collecting and analyzing large volumes of transactional data (Dai & Vasarhelyi, 2017 [10];
Issa et al., 2016 [8]). Therefore, it is argued that AI could be useful in audit processes such
as materiality and risk assessment, control assessment, audit planning, opinion selection,
and reporting (Bierstaker et al., 2014 [11]). Other benefits identified in the literature include
reducing human error (Murphy, 2017 [12]), facilitating continuous auditing (Brennan et al.,
2017 [13]), and the ability to audit all transactions, as well as reducing the cost and time
required for auditing (Issa et al., 2016 [8]; Westhausen, 2016 [14]).

2. Literature review
There are several perspectives on the definition of AI. The most important definitions
belong to authors who have studied the field in the last decade. Colom et al. (2008) [15]
defines AI in the context of problem solving, reasoning and learning, while Munoko et al.
(2020) [16] defines it as a new technology that resembles and reproduces human cognitive
abilities and judgments. Hassani et al. (2020) [17] describe it as intelligent systems designed
for data analysis and decision making, supporting the generation of results and insights
Pag. 139 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

from the analysis of voluminous and complicated data. AI is defined as a technology that
attempts to replicate or imitate human cognitive abilities, including judgment and
reasoning. With the advancement of the fourth industrial revolution, the use of AI
technologies has become increasingly common in various fields, such as education,
security, health, and including accounting and auditing (Mhlanga, 2021) [18]. At the same
time, AI is defined as the application of information systems and engineering with
intelligent machines and computers capable of exhibiting human traits of reasoning,
learning, and autonomous action, and analyzing big data and making quality decisions.
Hasan (2021) [19] describes AI as a form of rare intelligence manifested by machines or
robots, which perceive the environment and act to maximize their chances of achieving
their goals, based on programming and commands received.
Allami (2022) [20] provides a nuanced perspective on the multifaceted environment of AI
by summarizing the defining characteristics of this concept. These encompass not only
basic aspects such as perception, decision-making and prediction, but also more complex
functions such as automated information extraction, interactive communication, logical
thinking and the dynamic process of machine learning. AI aims to replicate and enhance
human cognitive capabilities, helping companies see and understand their environment
using digital computers or computer-controlled machines. Over the years, AI applications
have become increasingly relevant in a variety of social and economic areas, including
public health, transportation, education, security, communications, and defense.
Autonomous algorithms are currently driving progress in this field and are having a
substantial impact across numerous industries. In short, AI, developed by humans, has
evolved to a point where it is effortlessly integrated into everyday life and business.
Characteristics included in this context are the use of data to guide operations, the ability to
understand and interact with people, adaptability to change, and enhancement of human
capabilities (Figure 1.):

Pag. 140 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Improves financial reporting by enabling the rapid replication

of human intelligence, knowledge, and awareness in a
programmed computer.
• Mimics and displays cognitive skills associated with learning
and problem-solving.
• Represents a tool for the logical extraction of data and
providing accurate forecasts.
•Contributes to the automation of accounting processes and
risk detection in data sets.
•Enhances audit quality by automating accounting tasks and
saving time.
•Allows accountants to focus on advisory roles and strategic
decision-making.
• Includes components such as neural networks, genetic
algorithms, and natural language processing.
Characteristics of AI • Neural networks simulate the structures of the human brain
and facilitate machine learning.
• Genetic algorithms use natural selection and evolution to find
solutions to complex problems.
•Contributes to the optimization of processes and innovation in
various fields, including accounting..
•Enables communication between AI systems using human
natural language.
• Improves audit quality and the credibility of financial
reporting.
• New technologies disrupt existing structures and may replace
outdated ones.
• Technological innovations allow new businesses to compete
with established firms and introduce new ways of conducting
operations.
Figure 1: Characteristics of AI5

In a strict definition, AI represents the imitation of human intelligence by computers.

However, purists emphasize that many current applications are still relatively simple and,
therefore, cannot be considered true AI. This observation makes the definition inadequate,
as it would suggest that AI does not exist today. A common definition considers AI as a
technology that enables machines to imitate various complex human abilities. However,
this definition remains vague without specifying these "complex human abilities."
A similar definition was presented by the High-Level Expert Group on Artificial
Intelligence (AI HLEG) of the European Commission (EC): "Systems that exhibit
intelligent behavior by analyzing their environment and taking actions – with a certain
degree of autonomy – to achieve specific goals." (European Commission, 2018). The
definition provided by the AI HLEG encompasses all applications that we currently classify
as AI, while also leaving room for future changes to that classification. The defining
elements of the two fundamental areas of AI are presented in Figure no.2.

5 Source: Author conception based on the review of the specialized literature.

Pag. 141 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 2: Subdomains of AI6

3. Research Methodology
In conducting the study, both qualitative and quantitative analyses were used. The
arguments for using both approaches were: i) Holistic perspective: The qualitative analysis
provided an in-depth understanding of the issues and the context in which AI technologies
are applied in auditing. It allowed for exploring the perspectives and perceptions of industry
experts, as well as identifying unforeseen aspects and problems that could affect audit
quality; ii) Exploring complexity: Auditing and the application of AI in this field involve
complex aspects such as ethics, regulations, technical precision, etc.; iii) Identifying
emerging trends and patterns: Quantitative analysis was useful for identifying emerging
trends and patterns in audit data and AI usage. This provided solid evidence to support
certain conclusions and the development of research hypotheses; iv) Validating
conclusions and generalizing results: The use of both qualitative and quantitative analysis
helped validate the conclusions and ensure that the results obtained are robust and can be
generalized within the scientific community and in practice; v) Multidisciplinary
approach: A comprehensive study of AI application in auditing required a multidisciplinary
approach. Using both qualitative and quantitative analysis allowed for the integration of
perspectives from multiple fields, such as computer science, accounting, auditing,
organizational psychology, etc.; vi) Assessing impact and effectiveness: Quantitative
analysis was used to evaluate the impact and effectiveness of AI in auditing by measuring
performance indicators and comparing the results with traditional audit methods.

4. The Multiple Facets of Artificial Intelligence in Auditing

In recent years, advances in IT, particularly in the field of AI, have had a significant impact
on the accounting and auditing industry. These professions have undergone fundamental
changes as a result of progress in cognitive machine technologies, with a focus on the

6 Sursa: Ongsulee, P. (2017). Artificial intelligence, machine learning and deep learning. 2017 15th
International Conference on ICT and Knowledge Engineering (ICT&KE)
Pag. 142 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

development and application of AI. Dunn and Hollander (2017) [21] focus on the influence
of AI on auditing, redesigning AI system development based on the identified benefits and
limitations. This research explores how AI can enhance the effectiveness and quality of the
auditing process. The findings indicate that auditing firms, especially large ones, will
continue to invest in specialized expert systems and neural networks tailored to the industry
and specific audit tasks in order to minimize audit risks (Bogdan et al., 2023 [22]).
Additionally, large multinational corporations can develop their audit functions to use such
systems and strengthen internal control systems while reducing business risks. AI
technology allows them to manage large volumes of data, identify anomalous transactions,
and assess risks (Dincă et al., 2023 [23]).
The auditing profession is guided by International Standards on Auditing (ISA). According
to ISA 200, the application of AI in auditing represents the replication of human intelligence
functions by machines in performing the audit function. The general objective of a financial
statement audit performed by an independent auditor is to carry out the audit in accordance
with these standards. The auditor's objective, as per ISA 200, is to obtain reasonable
assurance to express an opinion regarding the absence of fraud and errors in the financial
statements and to issue an audit report communicating the audit findings. According to
Baldwin et al. (2006) [2], AI can be applied in auditing to perform a series of specific tasks
(Figure 3).

Figure 3: Specific Audit Tasks Suitable for AI Utilization7

The integration of data analysis into audits represents a giant leap in efficiency. Traditional
audit processes, which often rely on manual examination of financial records, are now
supported by sophisticated algorithms capable of processing vast data sets with remarkable
speed and accuracy. Auditors can use data analytics tools to identify patterns, anomalies,
and trends, allowing a more focused and targeted approach to the audit process. Data

7 Source: Baldwin, A.., Brown, C.E., & Trinkle, B.S. (2006). Opportunities for Artificial Intelligence
Development in the Accounting Domain: The Case for Auditing. Intelligent Systems in Accounting, Finance
and Management, 14(3), 77-86. https://doi.org/10.1002/isaf.277
Pag. 143 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

analysis not only accelerates the audit process but also allows auditors to explore the data
in-depth, revealing insights that would otherwise remain hidden in a manual review. This
enhanced efficiency results in time savings, enabling auditors to allocate resources more
strategically and focus more on areas with inherently higher risks.
Auditors are increasingly using continuous audit techniques, powered by data analysis and
blockchain technology. Continuous auditing enables real-time monitoring of financial
transactions, reducing the gap between the occurrence of an event and its detection.
Additionally, auditors are adapting to the dynamic nature of technology by integrating IT
audit skills into their toolkit. Assessing and understanding internal controls over
information systems becomes essential as technology increasingly blends with business
processes. Despite the rapid advancement of AI technologies in other areas, their adoption
in the audit and accounting profession has been slower. This is surprising, given the nature
of the field for applying AI technology, due to the various audit functions. However, there
are clear signs that AI adoption is gradually increasing in the audit and accounting
profession (Kokina et al., 2021 [24]).

5. Possible Solutions for Optimizing Audit Processes through the Use of Advanced
Technologies
Optimizing audit processes through the implementation of advanced technologies involves
utilizing the following elements (Figure no. 4):

Figure 4: Elements/technologies used for optimizing audit processes8

Continuous Auditing is defined as the systematic process of collecting electronic audit

evidence, providing a solid basis for issuing an opinion on the fair presentation of financial
statements in a real-time electronic accounting system. Continuous auditing is also
considered a detailed form of electronic auditing, allowing auditors to provide a certain
level of assurance regarding data as they are disclosed or immediately after disclosure. Zhao
et al. (2004) [25] emphasized that continuous auditing is associated with electronic
accounting systems, facing significant technical obstacles, lack of standards and guidelines,
the increased value of real-time financial information, and timely audit reporting.
Continuous auditing can generate either an "evergreen report" or an "on-demand report,"

8 Source: Authors' design

Pag. 144 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

and its use cases can cover all three professional services typically provided by independent
auditors – assurance, attestation, and audit services.
Decision Support Systems (DSS) are interactive, adaptable, and versatile software
platforms that assist in decision-making processes. These systems are designed to handle
structured management problems to enhance the decision-making process.
Neural Networks (NN) are a machine learning system that mimics the organization of the
human brain, composed of neurons and connections, and has the ability to adjust its
structure to perform tasks learned more efficiently. When neural networks become more
complex and include multiple layers, the term "deep learning" can be applied. Baldwin et
al. (2006) [2] and Deloitte (2018) examined the application of neural networks in the
Analytical Review Procedure used by auditors to obtain audit evidence.
Natural Language Processing (NLP) is a research field focusing on developing and using
artificial models to understand and process human language in a manner similar to humans
(Deloitte, 2018). Applications of NLP include processing unstructured textual information,
searching and analyzing documents automatically and systematically, as well as identifying
high-risk cases that deviate from preset targets. Fuzzy Logic is a reasoning technique that
mimics human thinking by allowing the evaluation of truth degrees of variables, with values
ranging from 0 to 1 (Baldwin et al., 2006) [2]. This approach allows handling the concept
of "partial truth" or "degrees of truth," better reflecting the complexity of the real world.
Fuzzy logic is used in areas such as assessing the risk of managerial fraud and making
significant decisions involving qualitative issues. Genetic Algorithm is a search method
inspired by the theory of evolution, where the best-adapted individuals are selected to
reproduce and pass on their traits to offspring. In the field of computing, genetic algorithms
use biologically inspired operators such as mutation, crossover, and natural selection to
develop efficient solutions to optimization and search problems. These algorithms are
effective in solving problems like transaction and account classification (Baldwin et al.,
2006 [2]). Genetic algorithms are also suggested for modeling auditor behavior in making
fraud-related decisions. Other applications of this algorithm include predicting bankruptcy
and making decisions regarding business continuity. Hybrid Systems are more suitable in
audit tasks that involve both quantitative analysis and qualitative judgment. These hybrid
IA technology systems combine various IA technologies, such as neural networks, fuzzy
logic, and genetic algorithms, to offer complex and adaptable solutions to the specific
requirements of audit tasks. Digital Audit represents a significant evolution of the auditing
process, characterized by the use of technology and digital data to perform and improve
audit procedures. This modern form of auditing relies on several defining characteristics
that give it uniqueness and efficiency in today’s digital business environment: i) Use of
technology, ii) Process automation, iii) Extensive data analysis, iv) Real-time data
monitoring, v) Data security and confidentiality.

6. Challenges of Integrating Artificial Intelligence in Audit Processes: Overcoming

Technical and Ethical Obstacles, Managing Risks
The review of specialized studies reveals that there are several significant challenges
associated with the integration of AI in auditing (Figure no. 5):

Pag. 145 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 5: Significant challenges associated with the integration of AI in auditing9

i) Data Quality and Integrity: The accuracy and reliability of AI algorithms critically
depend on the quality and integrity of the data fed into the system. Incomplete, inaccurate,
or biased information can distort results and lead to erroneous conclusions. Ensuring data
quality, implementing appropriate data governance frameworks, and establishing data
validation processes are essential to mitigate this challenge (Srinivasan & de Boer, 2020
[26]); ii) Ethical Considerations: The use of AI in auditing raises ethical concerns,
especially when dealing with sensitive financial and personal data. Protecting privacy, data
security, and compliance with relevant laws and regulations is imperative. Auditors must
consider ethical guidelines and establish protocols to protect data confidentiality and
maintain stakeholder trust; iii) Interpretation of Results: AI algorithms generate results
based on complex models and mathematical algorithms. Understanding and interpreting
these results can be challenging for auditors, particularly in the absence of adequate
expertise in AI and data analysis. Proper training and skill development for auditors are
necessary to correctly interpret and utilize information generated by AI (La Torre et al.,
2021 [27]); iv) Absence of Human Judgment and Professional Skepticism: AI systems
rely on predefined rules and algorithms, which may limit their ability to apply skepticism
and professional reasoning. Auditors bring valuable experience, intuition, and critical
thinking to the audit process, aspects that AI cannot fully replicate. Maintaining a balance
between AI-driven automation and human reasoning is essential for the integrity of the
auditing profession; v) Technical Complexity and Implementation Costs: Implementing
AI technologies in auditing requires specialized knowledge, infrastructure, and resources.
The initial costs of acquiring and implementing AI systems, as well as ongoing maintenance
and updates, can be significant. Small and medium-sized audit firms may face difficulties
in adopting AI due to the costs and technical complexities involved; vi) Regulatory and

9 Source: Own design

Pag. 146 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Compliance Considerations: The use of AI in auditing must comply with relevant

regulations and standards. However, the rapid pace of AI technology development often
outpaces regulatory frameworks, creating uncertainties and challenges. Auditors must stay
updated on regulatory developments and ensure that AI systems comply with applicable
legal requirements; vii) Aversion to Change: The adoption of AI in auditing may face
resistance from auditors and other stakeholders who are attached to traditional audit
practices. Concerns about job losses and the reliability of AI systems can slow down the
process of adopting and accepting AI in the auditing profession. Effective management of
these concerns, through change management, training programs, and clear communication,
is crucial.
According to Rebstadt et al. (2022) [28], while the importance of AI in auditing is evident
both academically and practically, the challenges associated with applying AI technologies
in the auditing profession remain largely unexplored. These challenges impact audit
outcomes, the decision-making process, and audit quality, raising ethical implications as
well. Despite the advantages of using machine learning algorithms in auditing, auditors may
fail to notice certain weaknesses associated with them, such as the "overfitting"
phenomenon. This refers to a situation where auditors cannot identify data features that do
not reflect real-world patterns. Additionally, machines may struggle to recognize that
statistically significant correlations between variables do not always indicate causal
relationships. These weaknesses highlight the potential challenges in results, emphasizing
the continued need for human evaluation and critical judgment in the audit decision-making
process, based on machine learning algorithm outputs. The predictive reliability of the
results from machine learning algorithm processing is closely tied to the quality of input
data, system design, methods used, and interpretation of output information. Thus, the
quality of the evidence generated through the use of these algorithms may be affected.
According to Gao & Han (2021) [29], implementing AI in auditing generates significant
changes in the audit process, influencing how evidence is generated and its quality. Audit
evidence should be generated from comprehensive and independent processes, such as
expert opinions, because investors rely on financial statements to make informed economic
or investment decisions. To make informed decisions, it is essential that financial statements
are audited, and managers provide accurate and reliable information to shareholders and
investors, who depend on the auditor's opinion to ensure the reliability of the information
provided. This assurance refers to the certainty and credibility of financial statements, based
on sufficient and appropriate audit evidence, according to ISA 500. Traditionally, auditors
have manually gathered this evidence, but the deficiencies of manual systems have often
affected the quality of evidence and audits in general.
The use of AI to collect, evaluate, and process large amounts of information from internal
and external environments can provide ample and diverse evidence. According to Gao &
Han (2021) [29], AI usage could improve the quality of audit evidence and reduce the gap
between audit expectations and reality, thereby altering the objective of the audit. They
suggest that the auditing profession could consider revising the ISAs to align them with the
use of AI in auditing.
Using artificial intelligence to collect information from independent sources can improve
the effectiveness and objectivity of confirmation evidence, separating audit procedures
from accounting processes. Even with the use of AI, there is recognition that an incorrect
opinion may be issued. Gao & Han (2021) [29] emphasize that when auditors make
Pag. 147 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

deductions, they may introduce their own subjective perspectives, individual differences,
and thought inconsistencies, which can lead to discrepancies between the audit conclusion
and reality or even result in incorrect opinions being issued. AI-assisted comprehensive
inferences, based on a rational model, can reduce the subjectivity of practitioners and make
their judgment less subjective and harder to contest.
The application of audit standards in the context of AI presents several challenges,
especially regarding fairness and transparency. There are issues related to what is measured
with the help of AI and the inputs used, without violating the rights and dignity of those
using AI technologies, even though these can contribute to correcting deficiencies in sample
representation. The use of analytical procedures involves certain assumptions according to
ISA guidelines, but machine learning has a distinct power in identifying specific
relationships or unexpected trends. Therefore, a revision of ISA standards and their
continued applicability in their current form for the auditing profession may be necessary,
considering the adoption of AI technologies. ISACA (2018) states that the lack of clear
audit standards assimilating emerging technologies affects the effectiveness and acceptance
of AI technologies.
Along with the opportunities AI offers, its application also brings a significant volume of
threats in the field of auditing and accounting. One of the main issues causing difficulties
is the regulatory environment. A study conducted by Deloitte in 2018 highlights the
regulation of cloud-based services, which varies globally, with certain European
jurisdictions imposing stricter restrictions. This discrepancy may give companies in less
restrictive jurisdictions an advantage in developing artificial intelligence technologies.
Apart from issues related to financial regulation, there are also other major risks. The same
study by Deloitte consulting highlights the dangers related to the complexity of financial
connections at both domestic and cross-border levels, the polarization of communities
around the development of artificial intelligence, and the risk of regional conflict or
financial exclusion among different population segments. In addition to these, there are
general threats associated with AI, such as job reduction and income inequalities caused by
the concentration of market power in the AI industry.
Auditors carefully examine the records and financial statements of an organization to
determine if they present a true and fair view of its financial position, performance, and
cash flows, in accordance with applicable accounting standards and regulations. They
follow a set of predefined procedures and standards to collect evidence and assess financial
information, which usually involves analyzing financial statements, verifying supporting
documentation, interviewing key personnel, and conducting tests and analyses to detect
errors, fraud, or non-compliance (Knechel & Salterio, 2016 [30]). The scope of the audit
extends beyond the financial statements, so auditors may also assess an entity's internal
control systems to evaluate the effectiveness of its internal processes and procedures for
financial reporting and risk management (Korol et al., 2022 [31]). They may also provide
recommendations for improving internal controls and mitigating risks (Knechel & Salterio,
2016 [30]). The introduction of AI in auditing generates significant implications for auditors
and the auditing profession as a whole (Figure no. 6).

Pag. 148 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 6: Implications of AI for Auditors10

An analysis of each challenge is as follows: i) Evolution of the skill set: The integration of
AI into auditing requires auditors to develop new skills and expertise. They must gain a
solid understanding of AI technologies, data analysis, and programming in order to
effectively use AI tools and interpret the results. Investment in continuous professional
development and skill enhancement programs is necessary to ensure auditors remain
competent and adapted to the AI-based audit environment; ii) Changing roles and
responsibilities: With the automation of repetitive tasks through AI, the roles and
responsibilities of auditors will undergo changes. As certain manual tasks become
redundant, auditors will focus more on value-added activities such as data analysis, risk
assessment, and providing strategic insights to clients. Adapting to and embracing their
evolving roles as trusted advisors and strategic partners for clients will be essential for their
success in this changing environment; iii) Increased efficiency and productivity: AI can
automate repetitive and time-consuming tasks, allowing auditors to be more efficient and
productive. This increase in efficiency can lead to faster audit cycles, more efficient
resource allocation, and an enhanced ability to manage larger volumes of data. This enables
auditors to focus their efforts on higher-value tasks that require professional judgment and
critical thinking; iv) Improved audit quality: The use of AI technologies can lead to an
improvement in audit quality by enhancing accuracy, identifying anomalies, and detecting
patterns in large datasets (Noordin et al., 2022 [32]). AI algorithms can process large
amounts of data quickly and consistently, reducing the risk of errors and oversight. This
leads to more reliable audit findings, improved risk assessments, and an overall
improvement in audit quality; v) Ethical and professional considerations: The use of AI in
auditing raises ethical concerns, such as confidentiality, data protection, and potential bias.
Auditors must ensure that AI systems are transparent, explainable, and comply with ethical
standards (Munoko et al., 2020 [16]). At the same time, skepticism and professional
judgment remain essential for addressing any limitations or biases that may arise in AI
10 Source: Original design based on the study of specialized literature.
Pag. 149 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

algorithms; vi) Collaboration with data specialists and AI experts: Auditors may need to
collaborate with data specialists and AI experts to effectively implement and utilize AI
technologies in auditing. This collaboration can facilitate the integration of AI into audit
processes, knowledge sharing, and ensure a multidisciplinary approach to audit
engagements (Noordin et al., 2022 [32]); vii) Impact of regulation and standardization:
The introduction of AI into auditing may involve the development of new regulations,
standards, and guidelines specific to AI-based audits. Regulatory authorities and standard-
setting organizations must adapt to technological advances to ensure the appropriate and
responsible use of AI in auditing (Noordin et al., 2022 [32]).

7. The influence of artificial intelligence on audit quality.

Audit service quality refers to complete, reliable, and comparable data that ensures the
quality of financial statements and their ability to add economic value to decisions made by
stakeholders. The quality of audit services results from auditors with ethical values,
integrity, attitude, and professional skills, audit independence, experience, and sufficient
time to perform a detailed audit. Some characteristics of high-quality audit services include
the use of appropriate and specific disruptive technologies, such as artificial intelligence
(AI), in processing and reporting financial-accounting data within public entities. The
application of AI has become a growing innovation in financial reporting and auditing
globally. The increase in the volume of corporate and business transactions has necessitated
the use of information technologies, and disruptive technologies now occupy a central place
in this field.
Recently, the value and quality of audit services have declined due to the increase in
financial reporting scandals globally (Noordin et al., 2022 [32]). Some have attributed the
complexity of audit quality to the use of manual computers, high levels of inaccuracy, and
delays in audit reporting. Moreover, recent research has highlighted favorable results and
economic benefits from the adoption and implementation of artificial intelligence. For
example, improvements have been reported in specific elements of AI, as shown in (Figure
no 7).

Figure 7: Benefits of adopting and implementing artificial intelligence11

11 Source: Abdollahi, A., Pitenoei, Y.R., & Gerayli, M.S. (2020). Auditor's report, auditor's size and value
relevance of accounting information. Journal of Applied Accounting Research, 21(4), 721-739.
https://doi.org/10.1108/JAAR-11-2019-0153 ; Greenman, C. (2017). Exploring the impact of artificial
intelligence on the accounting profession. Journal of Research in Business, Economics and Management, 8(3).
Pag. 150 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The successful application of AI can facilitate the understanding of historical data and the
prediction of data processing outcomes, avoiding information overload and ensuring the
accuracy and timeliness of financial reporting. Lee & Tajudeen (2020) [33] highlighted a
positive correlation between audit quality and the quality of financial reporting. While
numerous studies, including those by Chukwuani & Egiyi (2020) [1], have examined the
quality of the audit process from various perspectives, there is still a gap in the literature
regarding the investigation of AI’s effect on audit quality, particularly concerning the
implications for accountants, highlighting the need for further research due to
inconsistencies and the lack of clear conclusions in existing studies (Table no. 1).
Author Positive effects Negative effects Comments
AI is versatile and
flexible, contributing
Positive impact of
Hasan to the increased
information technology
(2021) reliability and
on audit qualityi
accuracy of financial
and audit reports.
Rezultate favorabile
Hemin privind efectul tehnologiei
(2017) informației asupra
calității auditului
Contradictory results
Lee &
regarding the impact of
Tajudeen
information technology on
(2020)
audit quality.
Negative impact of
Greenman disruptive technologies on
(2017) the credibility of audited
financial statements.
The increase in errors
due to the complexity of IT
systems, difficulties in
Balios & verifying data, and
colaboratorii excessive reliance on
(2020) technology that may
compromise the
professional judgment of
auditors.
AI recognized as a factor
Albawwat &
that can improve audit
Frijat (2021)
quality.

Table no. 1: Positive and negative effects of using information technology in performing
the audit.

Regarding financial reporting, AI represents a tool for the logical and structured extraction
of data, providing accurate and reliable forecasts. It contributes to improving the processing
and automation of document authorization to optimize internal accounting processes and
reporting. More specifically, AI uses computerized algorithms and programming to identify
and understand patterns and anomalies in datasets, allowing auditors to detect specific risk
areas more effectively and perform a variety of other audit and accounting processing tasks
at an unprecedented speed.
Disruptive technologies such as AI have revolutionized financial reporting processes,
replacing some of the conventional methods of financial reporting. New technologies have
brought clear benefits and profits to organizations that have adopted them over traditional
methods. Hasan (2021) [19] investigated the impact and implications of using AI in audits
on audit quality. The study used a structured questionnaire and an exploratory analysis of
Pag. 151 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

the relevant literature to examine multiple aspects of audit activities in which artificial
intelligence has been beneficial. The study concluded by highlighting benefits such as
accurate financial reporting, increased productivity, and auditor efficiency, compared to
traditional audit methods.

8. Case study regarding the application of specific AI mechanisms in internal auditing.

The quality of the activity carried out by the auditor can be measured by the product of their
work, the audit report, which is an essential element for supporting managerial decision-
making. In this sense, the case study conducted using the Orange Data Mining application
(Orange) aims to identify a relationship between input variables (number of internal
auditors/department, number of audit missions per year, number of recommendations per
total audit missions) and the output variable (the level of appreciation of the
recommendation) based on a regression equation and to identify the machine learning
technique that allows the best classification of test data. In this regard, a comparative
analysis will be made between logistic regression and the decision tree (DT) technique. DT
allows the division of a vast and heterogeneous collection of records into a series of
increasingly smaller and more homogeneous collections relative to a target attribute. The
mathematical foundation of logistic regression is represented by the Order Logit Model
with the following structure:

𝑃𝑟𝑜𝑏 𝑦𝑖 = 𝑗 𝑥, 𝑏, 𝑐 = 𝐹 𝑐𝑗 +1 − 𝑥𝑖 𝑏 − 𝐹(𝑐𝑗 − 𝑥𝑖 𝑏)
exp⁡
(𝑐𝑗 +1 − 𝑥𝑖 𝑏) exp⁡
(𝑐𝑗 − 𝑥𝑖 𝑏)
𝑃𝑟𝑜𝑏 𝑦𝑖 = 𝑗 𝑥, 𝑏, 𝑐 = −
1 + exp⁡
(𝑐𝑗 +1 − 𝑥𝑖 𝑏) 1 + exp⁡
(𝑐𝑗 − 𝑥𝑖 𝑏)
where:
• Xi = is the vector of explanatory variables (number of internal auditors/departments,
number of audit missions per year, number of recommendations per total audit missions);
• Yj = 1 ..... 4 represents the four alternatives for choosing the endogenous variable, the
level of appreciation of the audit recommendation.
• b = the vector of regression coefficients.
• c = the technical coefficients
In this case study, logistic regression is of the multinomial type due to the fact that the
dependent variable, Level of Appreciation, has three response options: level to be improved,
functional appreciation level, and critical appreciation level. The overall scheme of the
widget-type elements leading to the testing of classification methods is shown in (Figure
no. 8).

Pag. 152 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 8: Classification testing in the Orange Data Mining application12

A special element of the entire process of establishing a mathematical relationship to

characterize the historical data set and enable the prediction of the audit opinion is the
adherence to the conditions for applying the tests related to logistic regression and decision
trees. These preliminary conditions for applying the tests refer to verifying the existence of
extreme values (outliers) among the data for the independent variables (number of internal
auditors/departments, number of audit missions per year, number of recommendations per
total audit missions).
Outlier values are considered values greater than Q3 + 1.5 x IQR or values smaller than Q1
- 1.5 x IQR and can be visualized using a BoxPlot diagram. Below, the Boxplot diagrams
for the predictor-type variables used in the analysis are shown (Figures no. 9.1-9.3)

12 Source: Own design in defining the workflow.

Pag. 153 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure no. 9.1: BoxPlot diagram for the independent variable Number of auditors.

Figure no. 9.2: BoxPlot diagram for the independent variable Number of missions.

It can be observed that all the predictor variables have outlier values. In this situation, these
values must be removed in order to avoid distorting the statistical determinations related to
logistic regression. The removal of outlier values was performed using the Outliers
component, which allows for the determination and visualization of these values (Table no.
2). By removing these values, the data set is prepared for the application of specific ML
techniques.

Table no. 2: Visualization of outlier values.

The algorithm for constructing a decision tree using ID3 (Iterative Dichotomizer) starts
from a classified data set. Assuming that the elements (instances) of the data set have a
series of attributes whose values are known, the decision tree is generated in such a way
that, by traversing it for a new instance with a new set of attribute values, the class in which
that instance falls can be determined. The application of the decision tree construction
algorithm to the existing data set provides the following graphical representation (Figure
no. 10).

Pag. 154 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure no. 10: The decision tree related to the internal audit opinion.13

Traversal of the decision tree is done from the root to the leaf-type nodes and stops when
all filtering criteria have been analyzed. In our example, all analyses of the predictor-type
variables have been exhausted. The algorithm used to generate the decision tree stops IF
the number of audit missions > 12, the number of auditors > 2, and the number of
recommendations > 27, THEN the level of appreciation of the audit opinion is critical. This
algorithm allows us to classify a new set of data into a specific label related to the level of
appreciation of the audit opinion. By applying the specific logistic regression calculation
algorithm, the regression coefficients were determined, as presented in (Table no. 3).

13 Source: Own design.

Pag. 155 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Table no. 3: Coefficients of the logistic regression equation.

The logistic regression equations that allow labeling new instances as medium or low class
are the following:
CRITICAL level of appreciation of the recommendations = - 3.57 - 0.21 * Number of
auditors - 0.44 * Number of audited objectives + 0.54 * Number of recommendations (1)
LEVEL TO BE IMPROVED of the recommendation appreciation = - 0.75 - 0.21 * Number
of auditors + 0.02 * Number of audited objectives + 0.34 * Number of recommendations
(2)
It should be noted that the reference category is the high level of appreciation of the audit
opinion, in which case the regression equation does not need to be written. In this case, the
goal was to determine the appreciation of the audit opinion for an unclassified data set
consisting of three instances (Table no. 4), both through the decision tree algorithm and the
logistic regression algorithm. The procedure for determining the label for each instance is
automated and does not involve any calculations from the data analyst. In this case, the
label options for the three instances, both for the DT algorithm and the logistic regression
algorithm, are highlighted in (Table no. 5).

Table no. 4: The test data set

Table no. 5: The result of applying the calculation algorithms for the decision tree vs.
the logistic regression equation.

Pag. 156 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Table no. 6: The Test and Score option to determine the overall performance of the
classifiers

Level of
Number of audited Number of
Number of Auditors recommendation
objectives recommendations
appreciation
3 9 15 To be improved
4 18 60 Critical
2 10 18 To be improved

Table no. 7: Classification of the internal audit opinion based on the logistic regression
algorithm. Source: Own design.

The exemplified case study aims to apply specific AI mechanisms to automatically

formulate assessments of the results of an internal audit mission when new information is
available. The algorithm can serve as a tool for verifying the quality of the internal auditors'
work, considering the similarities with audit missions conducted over time by internal
auditors. Although the current capabilities of machine learning are limited, it excels at
routine tasks. Due to the large amount of data involved and the complexity of the activities
that need to be completed, machine learning has the potential to increase the efficiency and
quality of the internal audit process, with direct consequences on the auditor's productivity,
allowing more time for review and analysis, and a stronger focus on high-risk areas. The
quality of internal audit services is a concept that cannot be measured solely by focusing on
aspects such as incident/non-compliance reporting or strict adherence to procedures, but
also by evaluating the efficiency and effectiveness of the recommendations provided, the
impact on continuous improvement, and the added value that internal audit brings to the
entity. In this case, the combination of human resources and technology was the factor
leading to the qualitative assessment of internal audit services.

9. Conclusions
Regarding auditing, the use of artificial intelligence involves using technologies to improve
audit processes. This implies modifying the audit process, reorganizing audit functions, and
Pag. 157 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

updating skills across the profession to remain relevant through investments in technology,
training, and continuous professional development (CPD). AI can be applied in various
functions of the audit profession, including performing analytical review procedures, risk
assessment, applying algorithms, classification functions, evaluating significance,
judgments related to assessing business continuity, projections regarding company failure,
and evaluating internal controls. The application of AI in different audit functions is
accompanied by controversies concerning ethical considerations and audit quality. It can be
said that AI highlights advantages such as accuracy, objectivity, and speed, but it can also
draw attention to challenges related to bias and fairness.

References
[1] Chukwuani, V.N., & Egiyi, M.A. (2020). Automation of accounting processes: impact
of artificial intelligence. International Journal of Research and Innovation in Social
Science (IJRISS), 4(8), 444-449.
[2] Baldwin, A.., Brown, C.E., & Trinkle, B.S. (2006). Opportunities for Artificial
Intelligence Development in the Accounting Domain: The Case for Auditing. Intelligent
Systems in Accounting, Finance and Management, 14(3), 77-86.
https://doi.org/10.1002/isaf.277
[3] Mohammad, S.J., Hamad, A.K., Borgi, H., Thu, P.A., Sial, M.S., & Alhadidi, A.A.
(2020). How artificial intelligence changes the future of accounting industry. International
Journal of Economics and Business Administration, 8(3), 478-488.
https://doi.org/10.35808/ijeba/538
[4] Bizarro, P.A., & Dorian, M. (2017). Artificial intelligence: The future of
auditing. Internal Auditing, 5(1), 21-26.
[5] Kokina, J., & Davenport, T.H. (2017). The emergence of artificial intelligence: How
automation is changing auditing. Journal of emerging technologies in accounting, 14(1),
115-122. https://doi.org/10.2308/jeta-51730
[6] KPMG, 2018. AUDIT 2025: The future is now. [Online]. Disponibil la:
https://i.forbesimg.com/forbesinsights/kpmg_audit2025/KPMG_Audit_2025.pdf
[7] Chui, M., Manyika, J., & Miremadi, M. (2016). Where machines could replace
humans-and where they can't (yet). The McKinsey Quarterly, 1-12.
[8] Issa, H., Sun, T., & Vasarhelyi, M.A. (2016). Research ideas for artificial intelligence
in auditing: The formalization of audit and workforce supplementation. Journal of
emerging technologies in accounting, 13(2), 1-20. https://doi.org/10.2308/jeta-10511
[9] Sun, T., & Vasarhelyi, M.A. (2017). Deep Learning and the Future of Auditing: How
an Evolving Technology Could Transform Analysis and Improve Judgment. CPA
Journal, 87(6).

Pag. 158 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[10] Dai, J., & Vasarhelyi, M.A. (2017). Toward blockchain-based accounting and
assurance. Journal of information systems, 31(3), 5-21. https://doi.org/10.2308/isys-
51804
[11] Bierstaker, J., Janvrin, D., & Lowe, D.J. (2014). What factors influence auditors' use
of computer-assisted audit techniques?. Advances in Accounting, 30(1), 67-74.
https://doi.org/10.1016/j.adiac.2013.12.005
[12] Murphy, H. (2017). Auditing to be less of a burden as accountants embrace
AI. Financial Times, 18.
[13] Brennan, B., Baccala, M., & Flynn, M. (2017). How auditing will incorporate
AI. CFO Auditing, 2.
[14] Westhausen, H.U. (2016). The tech-savvy auditor: Effective use of audit technology
can enable audit departments to provide valuable insights. Internal Auditor, 73(3), 18-21.
[15] Colom, R., Abad, F. J., Quiroga, M. Á., Shih, P. C., & Flores-Mendoza, C. (2008).
Working memory and intelligence are highly related constructs, but
why?. Intelligence, 36(6), 584-606. https://doi.org/10.1016/j.intell.2008.01.002
[16] Munoko, I., Brown-Liburd, H.L., & Vasarhelyi, M. (2020). The ethical implications
of using artificial intelligence in auditing. Journal of business ethics, 167(2), 209-234.
https://doi.org/10.1007/s10551-019-04407-1
[17] Hassani, H., Silva, E. S., Unger, S., TajMazinani, M., & Mac Feely, S. (2020).
Artificial intelligence (AI) or intelligence augmentation (IA): what is the future?. AI, 1(2),
https://doi.org/10.3390/ai1020008
[18] Mhlanga, D. (2021). Artificial intelligence in the industry 4.0, and its impact on
poverty, innovation, infrastructure development, and the sustainable development goals:
Lessons from emerging economies?. Sustainability, 13(11).
https://doi.org/10.3390/su13115788
[19] Hasan, A. (2022). Artificial Intelligence (AI) in Accounting & Auditing: A Literature
Review. Open Journal of Business and Management, 10, 440-465. https://doi.org/
10.4236/ojbm.2022.101026
[20] Allami, F.A.J., Nabhan, S.H., & Jabbar, A.K. (2022). A Comparative Study of
Measuring the Accuracy of Using Artificial Intelligence Methods as an Alternative to
Traditional Methods of Auditing. World Economics and Finance Bulletin, 9, 90- 99.
[21] Dunn, C.L., & Hollander, S. (2017). The impact of artificial intelligence on auditing.
Current Issues in Auditing 11, A1–A12.
[22] Bogdan, V., Rus, L., Gherai, D.S., Florea, A.G., & Bugnar, N.G. (2023). A Streamline
Sustainable Business Performance Reporting Model by an Integrated FinESG
Approach. Sustainability, 15(24), 16860. https://doi.org/10.3390/su152416860
[23] Dincă, G., Netcu, I. C., & El-Naser, A. (2023). Analyzing EU’s Agricultural Sector
and Public Spending under Climate Change. Sustainability, 16(1), 72.
https://doi.org/10.3390/su16010072
Pag. 159 / 350
Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[24] Kokina, J., & Davenport, T.H. (2017). The emergence of artificial intelligence: How
automation is changing auditing. Journal of emerging technologies in accounting, 14(1),
115-122. https://doi.org/10.2308/jeta-51730
[25] Zhao, N., Yen, D.C., & Chang, I.C. (2004). Auditing in the e‐commerce
era. Information Management & Computer Security, 12(5), 389-400.
https://doi.org/10.1108/09685220410563360
[26] Srinivasan, A.V., & de Boer, M. (2020). Improving trust in data and algorithms in the
medium of AI. Maandblad voor Accountancy en Bedrijfseconomie, 94(3/4), 147-160.
https://doi.org/10.5117/mab.94.49425
[27] La Torre, M., Botes, V.L., Dumay, J., & Odendaal, E. (2021). Protecting a new
Achilles heel: the role of auditors within the practice of data protection. Managerial
Auditing Journal, 36(2), 218-239.
[28] Rebstadt, J., Remark, F., Fukas, P., Meier, P., & Thomas, O. (2022). Towards
personalized explanations for AI systems: designing a role model for explainable AI in
auditing. Wirtschaftsinformatik 2022 Proceedings. 2.
https://aisel.aisnet.org/wi2022/ai/ai/2
[29] Gao, Y., & Han, L. (2021). Implications of artificial intelligence on the objectives of
auditing financial statements and ways to achieve them. Microprocessors and
Microsystems. https://doi.org/10.1016/j.micpro.2021.104036
[30] Knechel, W.R., & Salterio, S. (2016). Auditing: Assurance and risk. Routledge.
[31] Korol, V., Dmytryk, O., Karpenko, O., Riadinska, V. O., Basiuk, O., Kobylnik, D.,
Mishchenko, T. (2022). Elaboration of recommendations on the development of the state
internal audit system when applying the digital technologies. Eastern-European Journal of
Enterprise Technologies, 1(13), 115. https://doi.org/10. 15587/1729-4061.2022.252424
[32] Noordin, N.A., Hussainey, K., & Hayek, A.F. (2022). The use of artificial
intelligence and audit quality: An analysis from the perspectives of external auditors in
the UAE. Journal of Risk and Financial Management, 15(8), 339.
https://doi.org/10.3390/jrfm15080339
[33] Lee, C.S., & Tajudeen, F.P. (2020). Usage and impact of artificial intelligence on
accounting: Evidence from Malaysian organisations. Asian Journal of Business and
Accounting, 13(1). https://doi.org/10.22452/ajba.vol13no1.8

Bibliography
Abdollahi, A., Pitenoei, Y.R., & Gerayli, M.S. (2020). Auditor's report, auditor's size and
value relevance of accounting information. Journal of Applied Accounting Research,
21(4), 721-739. https://doi.org/10.1108/JAAR-11-2019-0153

Pag. 160 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Aksoy, T., & Gurol, B. (2021). Artificial intelligence in computer-aided auditing

techniques and technologies (CAATTs) and an application proposal for auditors.
Auditing ecosystem and strategic accounting in the digital era (pp. 361-384). Springer,
Cham. https://doi.org/10.1007/978-3-030-72628-7_17
Al-Aroud, S.F. (2020). The Impact of artificial intelligence technologies on audit
evidence. Academy of Accounting and Financial Studies Journal, 24, 1-11.
Albawwat, I., & Frijat, Y. (2021). An analysis of auditors’ perceptions towards artificial
intelligence and its contribution to audit quality. Accounting, 7(4), 755-762.
Bai, G.H. (2017). Research on the Application and Influence of Auditing Artificial
Intelligence. DEStech Transactions on Social Science, Education and Human Science,
(eiem). https://doi.org/10.12783/dtssehs/eiem2017/16097
Deloitte, (2018). 16 artificial intelligence projects from Deloitte practical cases of applied
AI. [Online]. Disponibil la: https://www2.deloitte.com/content/dam/Deloitte
/nl/Documents/innovatie/deloitte-nl-innovatie-artificial-intelligence-16-practical-cases.pdf
Emetaram, E., & Uchime, H.N. (2021). Impact of Artificial Intelligence (AI) on
Accountancy Profession. Journal of Accounting and Financial Management, 7(2), 15-25.
EY, (2019). Multimillion-dollar investment in EY Blockchain Analyzer delivers new
upgrades for blockchain and cryptocurrency audit and tax services. [Online]. Disponibil
la: https://www.ey.com/en_gl/news/2019/04/multimillion-dollar-investment-in-ey-
blockchainanalyzer-delivers-new-upgrades-forblockchain-and-cryptocurrency-audit-and-
taxservices
Grewal, D.S. (2014). A critical conceptual analysis of definitions of artificial intelligence
as applicable to computer engineering. IOSR Journal of Computer Engineering, 16(2), 9-
13. https://doi.org/10.9790/0661-16210913
Kostić, N., & Tang, X. (2017). The future of audit: Examining the opportunities and
challenges stemming from the use of Big Data Analytics and Blockchain technology in
audit practice. Accounting and Finance Journal.
Lehner, O.M., Ittonen, K., Silvola, H., Ström, E., & Wührleitner, A. (2022). Artificial
intelligence based decision-making in accounting and auditing: ethical challenges and
normative thinking. Accounting, Auditing & Accountability Journal, 35(9), 109-135.
https://doi.org/10.1108/ AAAJ-09-2020-4934
Polak, P., Nelischer, C., Guo, H., & Robertson, D.C. (2020). “Intelligent” finance and
treasury management: what we can expect. Ai & Society, 35(3), 715-
726. https://doi.org/10.1007/s00146-019-00919-6
Shivram, V. (2024). Auditing with AI: a theoretical framework for applying machine
learning across the internal audit lifecycle. EDPACS, 69(1), 22-40.
https://doi.org/10.1080/07366981.2024.2312025

Pag. 161 / 350

Article’s total number of pages: 24
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

CHANGES IN BEHAVIOR OF UNIVERSITY STUDENTS AFTER COVID-19

SELF-RESTRAINT PERIOD

Shogo KOYAMA1
Takune SAKAUE2
Nobutaka SUZUKI3

Abstract
The purpose of this study is to examine how the lifestyle and consumer behavior of
university students changed during and after the period of self-restraint caused by COVID-
19. In recent years, university students’ lifestyles have changed considerably due to the
outbreak of COVID-19. However, self-restraint restrictions have been easing, and this study
investigates how students’ lifestyles and consumer behaviors have evolved under these
circumstances. The results show that, excluding time spent on online classes, students’ lives
have not changed drastically since the self-restraint period. Although the COVID-19
outbreak led to significant lifestyle adjustments, students have not fully reverted to their
pre-self-restraint habits. Two possible explanations for this lack of change are (1) the
discovery of online conveniences during the self-restraint period and (2) students becoming
accustomed to a routine established during that time.
Keywords: Behavior of University Students, COVID-19, Self-Restraint Period, Post-Self-
Restraint Period
JEL Classification: D10

1. Introduction
The purpose of this study is to determine the changes in lifestyle and consumer behavior
among college students during and after the relaxation of self-restraint due to the novel
coronavirus (hereafter, “COVID-19”).
In recent years, college students’ lifestyles have changed significantly as a result of COVID-
19. In FY2020, when a state of emergency was declared in Japan, almost all university

1Nagaoka University of Technology, Information and Management Systems Engineering, Japan,

shogo4933@gmail.com
2Nagaoka University of Technology, Information and Management Systems Engineering, Japan,

s233313@stn.nagaokaut.ac.jp, corresponding author

3Nagaoka University of Technology, Information and Management Systems Engineering, Japan,

nsuzuki39@kjs.nagaokaut.ac.jp
Pag. 162 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

lectures were moved from in-person to online. Subsequently, according to the Ministry of
Education, Culture, Sports, Science and Technology (MEXT, 2022) [1], 87.8% of lectures
in the first semester of FY2022 at 1,165 national, public, and private universities and
colleges of technology nationwide returned to a fully or almost fully in-person format.
Following the outbreak of COVID-19, Japanese society as a whole also increased its use of
online platforms for purchasing and other activities. According to the Statistics Bureau of
the Ministry of Internal Affairs and Communications (2022) [2], online purchases have
risen every year since March 2020, when the first state of emergency was declared.
It is important to analyze the specific changes in the lives and consumer behaviors of college
students—who have experienced major lifestyle changes, including returning to in-person
classes—so that we may consider the future of student life. Therefore, this paper explores
how university students’ lives changed during and after the relaxation of COVID-19 self-
restraint.
After reviewing previous studies in Section 2, Section 3 provides an overview of the
research, Section 4 presents the results of the quantitative analysis, and Section 5 discusses
the qualitative analysis. Section 6 provides a broader discussion, and Section 7 concludes.

2. Review of Previous Studies

Research on how COVID-19 changed college students’ lives and related attitudes has been
addressed in several studies. Ariki and Isaka (2021) [3] examined changes in the lives of
Japanese university and junior college students about one year after the first state of
emergency was declared, and the stress associated with these changes. Their results showed
that over 90% of students reported having fewer opportunities to interact with others and
going out less often, indicating that most students experienced lifestyle changes due to
COVID-19. Furthermore, more than 80% of the students found these two restrictions to be
“painful,” suggesting a psychological impact as well. Interestingly, participants who
reported low stress levels regarding COVID-19 frequently used the phrase “got used to it,”
implying that their perception of the pandemic was different from those reporting high
stress.
Suzuki et al. (2022) [4] conducted a survey of Aichi Prefecture residents aged 15 and older
as of October 2021 to understand their attitudes and behaviors in the wake of the corona
disaster. Their results showed that outings for shopping, hobbies, entertainment, walking,
and dining out had decreased compared to pre-pandemic levels. This decline in outings was
linked to both motivation for living and stress. These findings are consistent with Ariki and
Isaka (2021) [3], who observed that more than 80% of students reporting fewer outings felt
distressed by that limitation.

Pag. 163 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Regarding consumer behavior, Takeda et al. (2020) [5] analyzed changes in in-store
shopping before and after the first state of emergency. Their results indicated that, while the
percentage of shoppers had decreased during many time periods under the state of
emergency, it had largely returned to pre-pandemic levels by July 30 of the same year.
However, a new tendency emerged: consumers avoided crowded times. This shift suggests
that purchasing behavior was indeed altered by the pandemic. In terms of online shopping,
Ohata and Ujihara (2022) [6] examined usage changes before and after the pandemic among
residents in both the Tokyo metropolitan area and Okayama Prefecture. They found that
those in their 20s were the most frequent adopters of online shopping in both regions.
Kanai (2020) [7] compared online classes and face-to-face classes among university
students. The study noted that students’ preferences for one format over the other could not
be fully explained by attributes such as academic year or department, stressing the diversity
of each student’s situation and preferences. As a result, it is difficult to satisfy all parties—
students, faculty, and staff—regarding the classroom format.
Thus, various studies have been conducted on college students and young adults that
experienced the COVID-19 pandemic, and college students and young adults experienced
major life changes from its onset through the self-restraint period. In FY2022, self-restraint
measures were eased as many classes returned to in-person formats, suggesting life was
trending back toward pre-pandemic norms. However, research on how university students’
lives changed after self-restraint measures were eased appears limited. Hence, this study
focuses on the lives of college students after self-restraint was relaxed.

3. Research Outline
This study was based on the research of Shigeta et al. (2024). Shigeta et al. (2024)
investigated changes in college students’ consumer behavior from before the pandemic to
November 2020, during the COVID-19 self-restraint period. Their findings revealed an
increase in online purchases and a decrease in visits to brick-and-mortar stores after
COVID-19 first spread. Accordingly, the present study investigates college students at the
same university to examine changes in their lives and behavior after the relaxation of self-
restraint.

Period covered by this study

Our analysis covers the period from March 2020 to November 2022. Within this interval,
March 2020 to March 2022 is defined as the self-restraint period, and April 2022 to
November 2022 is defined as the period of relaxed self-restraint. The first declaration of a
state of emergency occurred in March 2020, and, as noted by MEXT (2022) [1], most
Japanese universities returned to an in-person format in April 2022. We designate this
switch to in-person classes as the point distinguishing the self-restraint from the relaxed
period.
Pag. 164 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Data
This study was conducted from July to November 2022 using both questionnaires and
interviews. The questionnaire was sent to 114 university students residing in Niigata
Prefecture, and 26 valid responses were received. The interview survey was conducted with
8 university students in online or face-to-face format.

4. Quantitative Research
This chapter asks questions based on the research question of what changes have occurred
in the lives of college students during the period of self-restraint due to COVID-19 and
during the period of relaxation of self-restraint and describes the results of the data obtained
from the questionnaire.

4.1 Questionnaire
We adopted seven questionnaire items from Shigeta et al. (2024). The questions consisted
of seven items: number of outings per week, number of times going out to physical stores
per week, weekly average of free time per day, weekly average of time spent on the Internet
per day, time spent on online shopping using a PC per week, time spent on mobile devices
using online shopping per week, time spent online through online classes per week. Shigeta
et al. (2024) set the time period for each questionnaire item to one month, while we set the
time period in this study to one week.

Questionnaire survey results

Question During the After (after

period of self- relaxation of relaxation)-
restraint self-restraint (during self-
restraint)

1 How many times do you go 4.52 5.27 0.75

out per week?

2 How many times per week 3.42 4.53 1.11

do you go to actual stores?

3 How many hours of free 6.85 6.85 0.00

time do you have per day
on average per week (not
including commuting to
Pag. 165 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

school, sleeping, classes,

part-time jobs, etc.)
4 How many hours per week 8.02 7.69 -0.33
on average do you spend on
the Internet per day?

5 How many hours per week 0.74 0.52 -0.22

do you spend online
shopping with a PC

6 How many hours per week 0.70 0.83 0.13

do you spend online
shopping with a mobile
device?

7 How many hours per week 11.73 3.44 -8.29

do you spend online
through online classes?

Table 1: Comparison of means from the results of the questionnaire.

during self-restraint after relaxation

Mean Standard Mean Standard t-value

deviation deviation
M M
SD SD
Number of times out 4.52 2.96 5.27 2.57 2.28*

Number of times I went 3.42 4.45 4.53 5.03 3.15**

to the actual store
Free time 6.85 2.67 6.85 3.18 0.00

Total time spent on the 8.02 4.30 7.69 4.15 -0.99

Internet

Time spent on PC for 0.74 1.05 0.52 0.53 -1.71

online purchases

Pag. 166 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Time spent on mobile 0.70 1.39 0.83 1.60 0.94

devices for online
purchases
Online class time 11.73 8.56 3.44 5.47 -4.85**

* p<.05 ** p<.01
Table 2: Changes in Consumer Behavior (Two-tailed t-test)

Table 1 shows the average results of the questions and responses regarding the changes in
daily life, and a comparison of the differences between the averages of the two periods
shows that the number of outings, trips to actual stores, and time spent using mobile devices
for online purchases each increased after the easing of the self-restraint. However, the
average changes, with the exception of visits to physical stores (about 1), were less than 1.
The total time spent on the Internet, which has been decreasing since self-restraint was
lifted, and the time spent using PCs for online purchasing was also less than 1. On the other
hand, the time spent in online classes dropped substantially by about 8 hours after self-
restraint measures eased, presumably because universities returned to face-to-face classes.
A two-tailed t-test (Table 2) confirms that differences were statistically significant for the
number of outings, visits to physical stores, and time spent in online classes.

5. Qualitative Research
This chapter describes the questions and results of the interviews with eight college
students.

5.1 Question Content

The interview survey questions were conducted as shown in Table 3.
Questions (17 questions in total)

1 How many times a week on average do you go out?

2 How many times a week, on average, do you go out other than to go to school?

3 What is the purpose of going out other than commuting to school?

4 How many times a week, on average, do you go out mainly for shopping?
5 Do you spend more time shopping on weekdays or holidays?

Pag. 167 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

6 How many hours of free time per day do you have on average per week?

7 Has the amount of time spent on shopping “increased” between the beginning of
self-restraint period and after it was relaxed?

8 Has your time spent on shopping “decreased” from the beginning of self-restraint
period to the time it was relaxed?

9 Has there been any change in the amount of time spent in physical stores between
the beginning of self-restraint period and after it was relaxed?

10 Has there been any change in time spent on online shopping between the
beginning of self-restraint period and after it was relaxed?

11 What do you buy in physical stores but not online?

12 What do you buy online but not in physical stores?

13 What criteria do you use to separate your shopping in physical stores and online?

14 What devices do they use for online shopping?

15 What are the reasons for question 14?

16 If you use a PC for question 14, when did you start using a PC to buy products?

17 If you have a PC, when did you start owning a PC?

Table 3: Interview Survey Questions

5.2 Interview Survey Results

Table 4 summarizes the interviews with Students A through H. Those that asked about two
periods, one during the self-restraint and the other during the relaxation of the self-restraint,
were described as “during the self-restraint/after the relaxation of the self-restraint” using
the “/” symbol.

A B C D E F G H

1 5.5/6. 7/8 5/7 2/6 times 6/6 times 1.5/3 1.5/3.5 5/5 times
5 times times times times
times

2 4.5/4 4/5 3/4 1/1 times 1/1 times 1.5/1. 1.5/3.5 2/2 times
times times times 5 times
times

Pag. 168 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3 Shop Part- Shopp Shopping, Shopping Club Shoppi Shopping,

ping, time ing, eating out activit ng, change of
going job, hospit ies, Eating pace
out shoppi al going out
for ng visits, out
fun hobbi for
es fun,
shoppi
ng

4 1.5/1. 1/1 1/3 1/1 times 1/1 times 1/1 1.5/1.5 2/2 times
5 times times times times
times

5 Week Week Week Weekday(d Holiday Week approxi Holiday

day day day aytime) → →Holida day mately →Holida
→ → → Weekday(n y → the y
Week Week Week ighttime) Week same
day day day day

6 3.5h/3 9h/9h 7h/5h 10h/6h 7h/7h 2.5h/2 6h/6h 5.5h/5.5h

h .5h

7 Time Com Com Commutin Time for Play Time to No

spent mutin mutin g time, games time go out change
comm g time g time preparation
uting to to for going
to school school out
schoo , time
l and away
in the from
lab home

8 time Onlin Home Home No Onlin Home No

at e time time, self- change e time, change
home prepar catering classe indepen
ation, s dent
house study
work

9 No No Increa No change No No Increas Increase

chang chang se change chang e
e e e

Pag. 169 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1 No No Decre No change Increase Decre Decreas No

0 chang chang ase ase e change
e e
1 Food, Food, Food, Food, Medicine Food, None Food,
1 Clothi Clothi Clothi Clothing, House Cartoon
ng, ng ng Appliances hold
Appli goods
ances

1 Book Bulk Home Luxury None None Product Clothing,

2 s, purch applia products, s not not
conta ase nces products availabl available
ct items, not e in in
lenses non- available physica physical
genui in physical l stores stores
ne stores
produ
cts

1 Basic Clothi Items Buy luxury Buy items Buy If it is Buy items
3 ally ng that I items that are consu sold in you want
physi and want online difficult mable a right
cal other to see because I to judge items physica away,
stores items in don't want by the in l store, such as on
to see purch person to carry layman's physic buy it the day
the ased are cash eye in a al in a they go
actual in purch physical stores, physica on sale, at
produ physic ased store; buy buy l store; a physical
cts, al at heavy long- if not, store.
and stores physic items and lasting buy it
online after al items that items online.
for online stores, spoil in online
items failure while temperatu
that s; consu re online.
canno items mer
t be with electr
purch many onics
ased altern are
in the atives purch
actual purch ased
stores ased online
. .

Pag. 170 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

online
.
1 smart smart smart smart PC PC PC smart
4 phone phone phone phone phone

1 Becau Becau Becau Because I To Becau Becaus Because it

5 se I'm se it is se it is can buy compare se I e I use is easy to
used easy easy from products can a PC all operate
to to to anywhere on PC look day
buyin chang operat things long
g e e up
quickl
y

1 High High Universit Unive High

6 school school y students rsity school
studen studen studen student
ts ts ts s

1 Unive High High University High Unive Univers Universit

7 rsity school school students school rsity ity y students
stude studen studen students studen student
nts ts ts ts s

Table 4: List of Interview Survey Results

Six of the eight students reported increased outings after self-restraint was eased, while two
reported no change and none reported a decrease. Three students (B, C, and G) increased
their frequency of going out for purposes other than commuting to school when self-
restraint was relaxed, while four respondents (D, E, F, and H) did not change their frequency
of going out, and one respondent (A) decreased. In addition, the number of times they went
out for the main purpose of shopping at actual stores remained unchanged except for C.
Even for C, the respondents had few opportunities to go out shopping because they had
gone home during the period of self-restraint. In addition, C also said that she had fewer
opportunities to go out for shopping because she had gone back home during the self-
restraint period. This is consistent with the findings during the quantitative study.
Regarding daily free time, which refers to time spent outside commuting to school,
sleeping, classes, and part-time jobs, three of the eight students (A, C, D) reported a
decrease, while the others saw no change and none reported an increase. In addition, five
of the eight respondents had an increase in time spent commuting to school and going out,
and a decrease in time spent at home. Due to this decrease in time spent at home, some
respondents were unable to do household chores in their spare time between classes, and
some said that they had more opportunities to eat out. Conversely, Student H did not change
how they spent their time between the two periods, indicating minimal lifestyle change.
Pag. 171 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Shifts in consumer behavior were also observed. Regarding the time spent in physical
stores, C, F, and H increased, while the other five respondents showed no change and none
decreased. With regard to shopping in physical stores, several students were unconcerned
about whether they shopped quickly or took their time. On the other hand, regarding time
spent on online purchases, three respondents (C, F, and G) decreased, four respondents (A,
B, D, and H) showed no change, and one respondent (E) showed an increase. The reasons
given by those who decreased were that they had more opportunities to go out and that they
had started to use actual stores again. As for E, whose time of use increased, they indicated
that they learned about the convenience of online purchasing as a result of their self-
restraint, and that their time of use is still increasing. The above results suggest that one of
the reasons for the decrease in time spent on online purchasing may be the number of times
the respondents go out.
We also conducted a survey to determine the criteria by which they use both physical stores
and online, and what they purchase. The results showed that the common reasons for
visiting physical stores included wanting to see products in person or being able to buy
items locally. Some respondents indicated that they would purchase products that they could
not judge in layman's terms or that they wanted right now at a physical store. On the other
hand, the criteria for using online varied from person to person. As examples, some cited
psychological reasons, such as buying luxury items online because they did not want to
carry large amounts of cash, and others cited high physical costs, such as buying items
online that are perishable in temperature or heavy because they do not have a car.
Finally, we asked about preferred devices for online shopping. The results show that two
respondents use only a smartphone, five use both a PC and a smartphone, and one uses only
a PC. The most common reasons for using a smartphone were ease of operation and the
ability to make purchases from any location.
On the other hand, reasons for using PCs include convenience, such as easy comparison
and research of products, and personal reasons, such as using a PC all day long. In sum,
among the eight people surveyed, a large percentage used smartphones, and their reasons
were largely in terms of operability.

6. Discussion
The results of the quantitative analysis in this study were significant with respect to the
number of outings, the number of visits to actual stores, and time spent in online classes,
based on the two-tailed t-test in Table 2. However, since the difference between the
meanings of the two periods was generally less than 1, this indicates that little change in
lifestyle occurred between the self-restraint period and after the relaxation of self-restraint,
apart from the time spent in online classes. Considering the findings by Ariki and Isaka
(2021), who showed that university students’ lives changed from before the outbreak of the
new coronavirus epidemic to the self-restraint period, it can be concluded that students’

Pag. 172 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

lives changed only during the self-restraint period and had not changed much since then, as
of November 2022. There are two possible reasons for this lack of change.
The first reason is that self-restraint due to COVID-19 may have prompted people to learn
about convenience. Ohata and Ujihara (2022) found that the frequency of online shopping
use increased most among those in their 20s before and after the pandemic. Comparing
these results with those of the present study, online shopping use rose during the self-
restraint period compared to before the pandemic, while it did not change much between
the self-restraint period and after the relaxation of self-restraint. In this regard, the
qualitative findings suggest that “convenience” may explain why lifestyles remained the
same, as individuals who increased their time spent on online shopping after restrictions
eased appear to have learned about its convenience and continued to use it.
The second possibility is that people become accustomed to a self-restraint lifestyle, and it
becomes routine. In the qualitative part of this research, only 3 out of 8 respondents
increased the number of times they went out for purposes other than commuting to school.
Compared to the results from Ariki and Isaka (2021), in which more than 80% of students
who went out less from April to July 2021 felt it was painful, it is possible that fewer people
felt the same way once the self-restraint was relaxed. Furthermore, the fact that those who
experienced low stress in Ariki and Isaka’s (2021) study used the phrase “got used to it”
suggests they may have adapted to a self-restraint lifestyle over time, resulting in little
change and a routine that persisted after restrictions were eased.

7. Conclusion
In this study, the research question was how college students, who were forced to make
major lifestyle changes as a result of self-restraint caused by the new coronavirus (COVID-
19), would subsequently alter their lifestyles and consumer behaviors once self-restraint
measures were eased.
The analysis showed that, unlike the pre-pandemic period and the self-restraint period—
when students' lives changed significantly, there were no notable changes in students' lives
from the self-restraint period to after the relaxation of self-restraint, apart from a reduction
in online class hours. This is the principal finding of this study.
There are two possible reasons for the lack of further change. The first is that the self-
restraint caused by COVID-19 may have prompted students to discover the convenience of
online activities. The second is that university students might have become accustomed to
living under self-restraint, such that this lifestyle became routine. However, these two
factors remain only possibilities.
Although this study discussed changes in college students' lives following the relaxation of
self-restraint and explored the factors influencing these changes, the small sample size in
both the quantitative and qualitative analyses poses limitations for validity and
Pag. 173 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

generalizability. Additionally, the impact of eased self-restraint measures needs ongoing

observation, indicating that more research is required.

Acknowledgments
We would like to express our sincere gratitude to everyone who participated in the
questionnaire and interview surveys for this study. This work was supported by JSPS
KAKENHI Grant Number 24K05055.

References
[1] MEXT. (2022). Daigaku tou ni okeru reiwa 4 nendo zenki no jugyou no jisshi houshin
tou ni kansuru chousa oyobi gakusei no shuugaku joukyou (chuutai ・kyuugaku) tou ni
kansuru chousa no kekka ni tsuite (shuuchi) [Results of the Survey on Class
Implementation Policies, etc. at Universities, etc. and the Survey on Students' Study
Status (Withdrawal, Leave of Absence, etc.) for the First Semester of the 2022 Academic
Year (Notification)]. https://www.mext.go.jp/content/20220603-mxt_kouhou01-
000004520_03.pdf (in Japanese).
[2] Statistics Bureau, Ministry of Internal Affairs and Communications. (2022). Kakei
shouhi joukyou chousa netto shoppingu no joukyou ni tsuite (futari ijou no setai) -2022
nen (reiwa 4 nen) 11 gatsu bun kekka- [Household Consumption Survey: Internet
Shopping (Households with Two or More Members) - November 2022 Results].
https://www.stat.go.jp/data/joukyou/pdf/n_joukyo.pdf (in Japanese).
[3] Isaka, Y., & Yuuki, E. (2021). Daigakusei ni okeru shingata korona uirusu kansenshou
ni yoru seikatsu no henka to sutoresu～sougo kyouchou teki jikokan to kankei ryuudousei
tono kanren [Lifestyle Changes and Stress Due to COVID-19 Among University Students
in Japan : Relations to interdependent-self construal, relational mobility, and morality].
Nichidai seikatsuka kenpou [ Report of the Research Institute of Science for Living, Nihon
University], (44),29-41 (in Japanese).
[4] Suzuki, Y., Yamazaki, M., & Mimura, Y. (2022). Korona ka ni okeru gaishutsu
katsudou to sutoresu ・ ikigai tono kankei ni kansuru kenkyuu – aichiken no seireishi ・
chuukakushi deno tyousa no kekka yori- [The Study on the Relationship between Going-
out Activities and Stress/ a sense of purpose in life in the COVID-19 -Results of a Survey
in the ordinance-designated cities and core cities of Aichi Prefecture-]. Toshi keikaku
ronbun syuu [Journal of the City Planning Institute of Japan], 57(3), 1164-1171 (in
Japanese).
Pag. 174 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[5] Takeda, R., Komatsuzaki, R., & Taniguti, M. (2020). COVID-19 ga motarashita
seikatsu henka no danryokusei – kinkyuu jitai sengen go 3 danmen deno daiari-de-ta wo
mochiite- [The elasticity of the changes of behavior due to COVID-19 -Using diary data
in three sections before and after the declaration of the state of emergency-]. Toshi keikaku
houkoku syuu [Reports of the City Planning Institute of Japan], 19(3), 311-317 (in
Japanese).
[6] Ohata, Y., & Ujihara, T. (2022). COVID-19 no ryuukou ni okeru netto shoppingu tou
no riyou henka no jittai haaku- toukyoutokubu to Okayama ken Okayama shi no
juukyosha wo taishou to shite- [The actual situation of changes in the using online
shopping during the COVID-19 Pandemic -For residents of Tokyo's wards and Okayama
city-]. Toshi keikaku ronbun shuu [Journal of the City Planning Institute of Japan], 57(1),
151-156 (in Japanese).
[7] Kanai, M. (2020). Onrain jugyou ni kansuru gakusei anke-to no gaiyou [Summary of
Student Survey on Online Classes]. dai 9 kai daigaku jouhou ・ kikan chousa kenkyuu
shuukai ronbun shuu [Proceedings of the Ninth Meeting on Japanese Insitutional
Research], 168-173 (in Japanese).
[8] Shigeta, T., Sakaue, T., & Suzuki, N. (2024). Changes in Consumer Behavior of
University Students during COVID-19 Self-Restraint Period. Journal of Information
Systems & Operations Management, 18(2), 185-198.

Bibliography
Asaba, S. (2002). Nihonkigyou no kyousougenri [Competitive Principles of Japanese
Companies]. Tokyo, Japan: Toyo Keizai Inc. (in Japanese).
Isaka, Y., & Yuuki, E. (2021). Daigakusei ni okeru shingata korona uirusu kansenshou ni
yoru seikatsu no henka to sutoresu～sougo kyouchou teki jikokan to kankei ryuudousei
tono kanren [Lifestyle Changes and Stress Due to COVID-19 Among University Students
in Japan : Relations to interdependent-self construal, relational mobility, and morality].
Nichidai seikatsuka kenpou [Report of the Research Institute of Science for Living, Nihon
University], (44),29-41 (in Japanese).
Kanai, M. (2020). Onrain jugyou ni kansuru gakusei anke-to no gaiyou [Summary of
Student Survey on Online Classes]. dai 9 kai daigaku jouhou ・ kikan chousa kenkyuu
shuukai ronbun shuu [Proceedings of the Ninth Meeting on Japanese Insitutional
Research], 168-173 (in Japanese).
MEXT. (2022). Daigaku tou ni okeru reiwa 4 nendo zenki no jugyou no jisshi houshin tou
ni kansuru chousa oyobi gakusei no shuugaku joukyou (chuutai ・kyuugaku) tou ni
kansuru chousa no kekka ni tsuite (shuuchi) [Results of the Survey on Class
Pag. 175 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Implementation Policies, etc. at Universities, etc. and the Survey on Students' Study
Status (Withdrawal, Leave of Absence, etc.) for the First Semester of the 2022 Academic
Year (Notification)]. https://www.mext.go.jp/content/20220603-mxt_kouhou01-
000004520_03.pdf (in Japanese).
Nomura, K. (2017). Shakaikagaku no kangaekata [Methods and Methodology in Social
Research]. Aichi, Japan: The University of Nagoya Press (in Japanese).
Ohata, Y., & Ujihara, T. (2022). COVID-19 no ryuukou ni okeru netto shoppingu tou no
riyou henka no jittai haaku- toukyoutokubu to Okayama ken Okayama shi no juukyosha
wo taishou to shite- [The actual situation of changes in the using online shopping during
the COVID-19 Pandemic -For residents of Tokyo's wards and Okayama city-]. Toshi
keikaku ronbun shuu [Journal of the City Planning Institute of Japan], 57(1), 151-156 (in
Japanese).
Shigeta, T., Sakaue, T., & Suzuki, N. (2024). Changes in Consumer Behavior of
University Students during COVID-19 Self-Restraint Period. Journal of Information
Systems & Operations Management, 18(2), 185-198.
Statistics Bureau, Ministry of Internal Affairs and Communications. (2022). Kakei shouhi
joukyou chousa netto shoppingu no joukyou ni tsuite (futari ijou no setai) -2022 nen
(reiwa 4 nen) 11 gatsu bun kekka- [Household Consumption Survey: Internet Shopping
(Households with Two or More Members) - November 2022 Results].
https://www.stat.go.jp/data/joukyou/pdf/n_joukyo.pdf (in Japanese).
Sukegawa, M., Hatsushio, K., & Suzuki, N. (2024). Reconsideration of Long-term
Internships and Future Proposals. Journal of Information Systems & Operations
Management, 18(1), 261-275.
Suzuki, Y., Yamazaki, M., & Mimura, Y. (2022). Korona ka ni okeru gaishutsu katsudou
to sutoresu ・ ikigai tono kankei ni kansuru kenkyuu – aichiken no seireisi ・
chuukakushi deno chousa no kekka yori- [The Study on the Relationship between Going-
out Activities and Stress/ a sense of purpose in life in the COVID-19 -Results of a Survey
in the ordinance-designated cities and core cities of Aichi Prefecture-]. Toshi keikaku
ronbun shuu [Journal of the City Planning Institute of Japan], 57(3), 1164-1171 (in
Japanese).
Takeda, R., Komatsuzaki, R., & Taniguti, M. (2020). COVID-19 ga motarashita seikatsu
henka no danryokusei – kinkyuu jitai sengen go 3 danmen deno daiari-de-ta wo mochiite-
[The elasticity of the changes of behavior due to COVID-19 -Using diary data in three
sections before and after the declaration of the state of emergency-]. Toshi keikaku
houkoku shuu [Journal of the City Planning Institute of Japan], 19(3), 311-317 (in
Japanese).

Yamauchi, Y. (2015).「Toso」toshiteno service [Service as a Struggle]. Tokyo, Japan:

Chuoukeizai-sha Holdings, Inc. (in Japanese).
Pag. 176 / 350
Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Yin, R.K. (1994). Case Study Research: Design and Methods (2nd edition). California,
America: Sage Publications, Inc.

Pag. 177 / 350

Article’s total number of pages: 16
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ARTIFICIAL INTELLIGENCE SOLUTIONS FOR ENERGY CONSUMPTION

OPTIMIZATION IN IOT DEVICES

Răzvan MOCANU1
Florentina NIDELCU2
George CĂRUȚAȘU3

Abstract
The rapid expansion of Internet of Things (IoT) devices has substantially increased global
energy demands, leading to critical economic and environmental challenges. Traditional
energy management techniques are increasingly inadequate, necessitating the integration of
sophisticated artificial intelligence (AI) solutions.
This study addresses the urgent issue of rising energy consumption driven by the
exponential growth of Internet of Things (IoT) devices. It proposes a hybrid artificial
intelligence (AI) methodology that integrates supervised machine learning and deep
reinforcement learning to optimize real-time energy usage in heterogeneous IoT
environments. Through extensive simulations and analysis of real-world case studies, the
proposed models demonstrate up to 30–40% improvements in energy efficiency compared
to conventional rule-based methods. The novelty of this research lies in its comparative
performance evaluation of multiple AI approaches across different IoT domains, offering a
replicable framework for smart building management, industrial IoT, and smart grids.
Keywords: Artificial Intelligence, IoT, Energy Optimization, Sustainability
JEL Classification: Q42, L86

1. Introduction
The accelerated adoption of IoT devices worldwide has significantly elevated energy
consumption, placing increased pressure on global energy resources. With IoT devices
projected to surpass 24 billion by 2030, traditional approaches to energy management are
no longer sufficient. The complexity and dynamic nature of IoT networks necessitate

1PhD Candidate Razvan Mocanu, National University of Science and Technology Politehnica Bucharest,
Romania, razvan_sorin.mocanu@stud.fiir.upb.ro, corresponding author
2PhD Candidate Florentina Nidelcu National University of Science and Technology Politehnica Bucharest,

Romania, florentina.nidelcu@stud.fiir.upb.ro
3Prof.PhD.Hab George Carutasu Romanian-American University, National University of Science and

Technology Politehnica Bucharest, Romania, george.carutasu@rau.ro

Pag. 178 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

innovative solutions that leverage artificial intelligence to enable predictive and adaptive
management strategies, fostering energy efficiency and sustainability.
Moreover, the Internet of Things (IoT) is revolutionizing various industries,
including smart cities, industrial automation, healthcare, and energy management.
However, its rapid adoption introduces significant concerns regarding energy efficiency and
sustainability. Traditional grid-based energy management systems struggle to accommodate
the increasingly decentralized and data-intensive nature of IoT networks. The need
for intelligent, data-driven, and real-time decision-making has led to a shift toward AI-
powered optimization techniques.
As a result, researchers have increasingly focused on integrating artificial intelligence
(AI) techniques, such as machine learning (ML), deep learning (DL), and reinforcement
learning (RL), to enhance energy efficiency within IoT ecosystems. These AI-driven
approaches can optimize energy consumption by predicting demand, dynamically
allocating resources, and minimizing unnecessary energy expenditure. Additionally, AI
facilitates the seamless integration of renewable energy sources, further promoting
sustainable energy usage. This paper investigates: To what extent can AI-based optimization
models reduce energy consumption in IoT systems while maintaining performance and
scalability?
Furthermore, the concept of Green IoT (G-IoT) has gained traction in recent years, aiming
to develop sustainable IoT solutions that reduce energy consumption and minimize
environmental impact. G-IoT incorporates energy-efficient sensors, cloud computing
strategies, and energy-aware communication protocols to improve overall system
performance while reducing the carbon footprint of IoT infrastructures.
This paper explores the convergence of IoT and AI-driven energy optimization techniques,
highlighting key methodologies, challenges, and real-world applications that can
enhance energy efficiency, reduce operational costs, and contribute to global sustainability
efforts.

2. Literature Review
Recent scholarly research has explored AI-driven solutions for energy optimization
extensively, demonstrating the potential of artificial intelligence in managing energy
demand and reducing inefficiencies across multiple sectors.
Studies highlighted how machine learning models can dynamically adjust energy
consumption patterns based on user behavior and environmental conditions, leading to
an average reduction of 20% in energy usage [5]. These findings align with prior studies
indicating that intelligent automation can minimize unnecessary power consumption in
residential and commercial buildings.

Pag. 179 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Other studies examined AI-powered IoT frameworks for energy efficiency in smart grids
and urban infrastructure [6]. Their research demonstrated that deep learning models,
particularly convolutional neural networks (CNNs) and long short-term memory (LSTM)
networks, can effectively forecast energy demand, detect anomalies in energy usage, and
optimize power allocation in real-time. The study confirmed that integrating AI-driven
forecasting techniques into grid management results in enhanced operational efficiency,
reduced peak loads, and improved renewable energy integration.
Similarly, another study underscored the importance of AI-based optimization in smart grid
systems[7]. Their research focused on the application of reinforcement learning (RL) for
adaptive demand-response mechanisms, allowing real-time energy distribution adjustments
based on supply and demand fluctuations. Their findings suggested that RL-driven
optimization models could reduce overall grid energy wastage by 15-25%, ensuring more
efficient use of renewable energy sources such as solar and wind power.
Another relevant study [3] investigated the impact of AI in industrial IoT (IIoT) settings,
where manufacturing and automation require precise energy management solutions. The
study found that predictive maintenance using AI-based anomaly detection
algorithms significantly reduced unplanned downtime, leading to a 30% decrease in wasted
energy associated with inefficient machinery operation.
A comprehensive review [4] on AI-driven energy optimization in IoT networks revealed
that AI can enhance energy-aware task scheduling and resource allocation in edge
computing environments. Their findings indicated that deep reinforcement learning (DRL)
algorithms enable real-time adaptation of IoT network configurations, leading to a 40%
increase in processing efficiency while reducing computational energy expenditure.
Moreover, a study [9] introduced the concept of Green IoT (G-IoT) as a framework for
integrating AI into sustainable energy solutions. This approach focuses on reducing
the carbon footprint of IoT systems by leveraging AI-based network optimization, energy-
efficient communication protocols, and adaptive power scaling techniques. The study
emphasized that the adoption of AI-enhanced cloud computing and energy-aware sensor
networks can significantly lower energy consumption in large-scale IoT deployments.
Collectively, these studies confirm that AI-driven optimization techniques have far-
reaching implications across diverse applications, from smart homes and industrial
automation to grid management and edge computing. The integration of machine learning,
deep learning, and reinforcement learning into IoT-based energy systems represents
a critical step toward achieving sustainable, efficient, and intelligent energy management.

3. Problem Definition
The pervasive integration of IoT devices across various sectors, including smart cities,
industrial automation, healthcare, and energy management, has significantly exacerbated
global energy consumption. Traditional energy management strategies struggle to handle
Pag. 180 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

the complexity, scale, and dynamic nature of IoT-generated data, making it increasingly
difficult to achieve real-time monitoring, predictive analytics, and adaptive optimization.
This challenge highlights the necessity for advanced AI methodologies that can
dynamically analyze large-scale data, optimize energy usage, and minimize environmental
impact.
3.1 Challenges in Energy Management for IoT Ecosystems
The rapid expansion of IoT presents significant challenges in energy management due to
high energy demand, vast data processing requirements, inefficiencies in traditional energy
systems, and difficulties in integrating renewable energy sources. The continuous operation
of billions of IoT devices leads to increased electricity consumption and carbon emissions,
while data-intensive networks require substantial computational power, contributing to
higher energy costs. Conventional energy-saving methods struggle to balance dynamic
demand and supply, and interoperability issues among heterogeneous IoT devices
complicate scalability.
Artificial intelligence (AI) offers a solution by enhancing predictive energy optimization,
real-time anomaly detection, adaptive scheduling, and distributed energy management.
Machine learning and deep learning models can analyze historical energy consumption data
to forecast demand and dynamically optimize power distribution. AI-based anomaly
detection systems improve efficiency by identifying abnormal energy consumption
patterns, preventing system failures, and enabling predictive maintenance. Smart
scheduling and workload allocation strategies powered by AI have shown up to 30% energy
savings in industrial IoT applications. Additionally, edge AI deployment reduces cloud
dependency and lowers energy consumption by up to 40%, while reinforcement learning
optimizes energy storage and distribution in smart grids, improving efficiency by 15-25%.
One of the critical issues is the inefficient allocation of energy resources due to the lack of
adaptive and predictive control mechanisms. Existing solutions fail to handle the non-linear
and time-dependent characteristics of IoT workloads, leading to energy wastage and poor
system performance. Figure 1 illustrates the key challenges in IoT energy management,
including unpredictable load variations, intermittent connectivity, and limitations in energy
harvesting from renewable sources.
This study aims to address the following research problem: How can AI models be designed
to adaptively and intelligently optimize energy consumption in diverse and dynamic IoT
environments?

Pag. 181 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 1: Key Challenges in loT Energy Management

The main limitations of this research include the availability and quality of real-time data,
the high computational requirements of deep learning models, and the difficulties of
deploying AI at scale in resource-constrained environments. These challenges will be
discussed in more depth in subsequent sections.

4. Development Methods and Algorithms

Recent advances in artificial intelligence have significantly enhanced the optimization of
energy consumption in IoT-based environments. Research studies emphasize the role of AI
in enabling real-time decision-making, predictive analytics, and adaptive energy
management [5][6].
The methodologies applied in this study include:
Machine Learning (ML): Regression models, clustering algorithms, and decision trees have
been widely employed for forecasting energy demands and detecting inefficiencies. Studies
[3] show that supervised ML models can achieve up to 85% accuracy in predicting energy
consumption trends, allowing better resource allocation in IoT networks.
Deep Learning (DL): Advanced DL techniques such as recurrent neural networks (RNNs),
long short-term memory networks (LSTMs), and convolutional neural networks (CNNs)
are extensively used in IoT energy management. Recent findings [4] demonstrate that DL
models trained on historical energy consumption data can reduce energy waste by 30% in
smart buildings by accurately predicting HVAC usage patterns.

Pag. 182 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Reinforcement Learning (RL): Reinforcement learning techniques have been applied for
dynamic and adaptive energy management. Studies [7] indicate that RL-based energy
optimization algorithms can improve smart grid efficiency by 25% by dynamically
balancing power loads in response to real-time demand and supply fluctuations.
Furthermore, hybrid AI methodologies combining ML, DL, and RL have proven effective
in industrial IoT applications, where real-time decision-making and energy-efficient
scheduling are critical [1]. The integration of AI-driven optimization strategies in IoT
infrastructures represents a significant leap toward achieving sustainable and intelligent
energy management solutions.

Figure 2. High-Level Flow Diagram of the Proposed AI-Driven Optimization Framework

Artificial Intelligence offers a robust toolkit for energy optimization in IoT systems through
methods such as machine learning (ML), deep learning (DL), and reinforcement learning
(RL). These approaches enable predictive analytics, adaptive control, and autonomous
decision-making, essential for managing the complex and dynamic nature of IoT
environments. Figure 2 presents a high-level flow diagram of the proposed AI-driven
optimization framework, detailing data acquisition, preprocessing, model selection,
training, and real-time deployment.

Pag. 183 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Machine Learning (ML): ML techniques, including regression analysis, decision trees, and
support vector machines, are effective in detecting patterns in historical energy usage and
forecasting demand. Their strength lies in interpretability and low computational cost.
However, they require extensive labeled data and perform suboptimally in highly dynamic
systems.
Deep Learning (DL): DL approaches such as convolutional neural networks (CNNs) and
recurrent neural networks (RNNs) can model complex, high-dimensional IoT data. Long
short-term memory (LSTM) networks are particularly useful for time-series forecasting of
energy consumption. DL models excel in feature extraction and predictive accuracy but
come with high training and computational costs.
Reinforcement Learning (RL): RL methods, especially deep reinforcement learning (DRL),
allow systems to learn optimal energy management strategies through interaction with the
environment. Techniques like Q-learning and policy gradient methods dynamically adjust
power consumption in response to changing conditions. RL excels in adaptability and
autonomy but often requires long convergence times and complex tuning.
Computational Complexity Discussion: ML models are lightweight and suitable for
deployment on low-power devices. DL models, while highly accurate, demand significant
GPU resources. RL approaches require the most extensive computational resources due to
their iterative, feedback-based learning structure. Choosing the right method depends on
the specific use case, hardware constraints, and required response time.
Together, these AI methodologies form the core of the proposed optimization framework,
supporting dynamic energy scheduling, anomaly detection, and demand forecasting across
diverse IoT applications.

5. Data Set Loading and Analysis

The success of AI-driven energy optimization in IoT environments depends significantly
on the quality and diversity of the datasets used. Recent studies have emphasized the
necessity of high-resolution, real-time datasets to improve the predictive capabilities of
machine learning (ML) and deep learning (DL) models [3][4].
5.1 Data Sources and Collection Methods
Comprehensive datasets were collected from IoT sensors deployed across various
environments to enable effective energy optimization. In commercial buildings, energy
monitoring systems were used to track power consumption, temperature fluctuations,
HVAC usage, and occupancy patterns, providing valuable insights into energy efficiency
[5]. In industrial settings, sensor-driven energy audits and predictive maintenance data were
gathered, including machine performance metrics and downtime analysis, to improve
operational reliability [6]. Smart grids benefited from real-time power demand-response
data, facilitating the integration of renewable energy sources and monitoring voltage
fluctuations for more efficient distribution [7]. Additionally, residential homes relied on
smart meters and IoT-enabled devices to capture real-time household energy consumption,
helping users improve energy efficiency and reduce costs [2].
Pag. 184 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

5.2 Data Preprocessing and Feature Selection

To ensure high-quality input data for AI models, the collected datasets underwent extensive
preprocessing, normalization, and feature selection. Data cleaning procedures were applied
to eliminate anomalies, missing values, and erroneous sensor readings, enhancing data
accuracy. Normalization techniques were used to standardize numerical values, thereby
improving training stability and facilitating model convergence. Feature engineering helped
identify critical variables such as peak energy usage hours, device-specific consumption,
and occupancy-based trends, which played a crucial role in optimizing predictive analytics
and decision-making processes [3].
5.3 AI-Driven Data Augmentation and Synthetic Data Generation
Given the limitations of real-world datasets, AI-based data augmentation techniques were
implemented to generate synthetic data, improving model training and overall predictive
performance. Generative adversarial networks (GANs) were particularly effective in
creating simulated energy consumption scenarios, which enhanced model robustness and
provided a broader dataset for deep learning applications [4]. These synthetic datasets
allowed AI models to generalize better across different IoT environments, reducing bias
and improving energy optimization outcomes.
5.4 Challenges in IoT Data Collection for Energy Management
Despite significant advancements in data collection, several challenges remain in IoT-based
energy management. Sensor calibration issues continue to affect data reliability, as
variations in sensor accuracy can lead to inconsistencies in energy monitoring. High data
transmission costs are another concern, as cloud-based data storage and real-time
processing significantly increase energy consumption. Additionally, privacy and security
risks are critical considerations, particularly in household and industrial energy
consumption monitoring, where strict access controls are necessary to prevent unauthorized
data breaches [6]
The processed datasets provided a strong foundation for training, validating, and deploying
AI-driven optimization models, ultimately enhancing the efficiency and adaptability of IoT-
based energy management systems. These advancements contribute to more effective
energy allocation, reduced consumption, and improved sustainability across various
sectors.

6. Software Computations and Implementation

Computational analyses for AI-driven energy optimization in IoT environments require
robust data processing frameworks, machine learning models, and real-time deployment
strategies. To achieve scalable, adaptive, and efficient energy management, multiple
computational approaches have been adopted in recent studies [4] [5].
6.1 Computational Frameworks for AI-Driven Energy Optimization
Pag. 185 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

AI-based energy optimization systems rely extensively on advanced machine learning (ML)
and deep learning (DL) frameworks to process large-scale IoT data efficiently. Several
computational tools were employed to support different aspects of energy management.
TensorFlow and Keras were utilized for deep learning model development, particularly in
implementing convolutional neural networks (CNNs) and recurrent neural networks
(RNNs), which are essential for predicting energy demand trends and optimizing HVAC
usage patterns in smart buildings [6]. Scikit-learn and XGBoost were applied in machine
learning tasks, including regression analysis, anomaly detection, and energy consumption
forecasting, enabling more accurate and data-driven decision-making [3]. PyTorch was
used in reinforcement learning (RL) algorithms to enhance adaptive energy optimization
and facilitate intelligent load balancing in smart grids, contributing to greater efficiency and
stability in energy distribution [7]. The selection of these computational frameworks was
based on their ability to handle model complexity, ensure scalability, and meet real-time
processing requirements across various IoT energy applications.

Framework AI Methodology Use Case

HVAC load prediction, time-series
TensorFlow/Keras Deep Learning modeling
Scikit-learn Machine Learning Regression, anomaly detection
Reinforcement
PyTorch Learning Smart grid power balancing
XGBoost Gradient Boosted Trees Energy usage classification
Table 1. Software Tools and Applications

6.2 Preprocessing and Feature Engineering

Data preprocessing plays a crucial role in improving model accuracy by ensuring high-
quality input data for AI-driven energy optimization. The process began with data cleaning,
which involved removing missing values, sensor errors, and inconsistencies to enhance the
reliability of machine learning models [4]. Following this, normalization and
standardization techniques were applied to adjust numerical features, ensuring
compatibility across various machine learning frameworks while reducing computational
overhead. Feature selection was then conducted to identify key variables, such as real-time
power consumption, environmental factors, and device operation schedules, which
significantly improve the predictive capabilities of AI models [5]. To further refine model
efficiency and reduce dimensionality, advanced feature extraction techniques, including
principal component analysis (PCA) and autoencoders, were implemented, enabling better
performance and faster processing of IoT-based energy data.
6.3 AI Model Training and Optimization
The training phase involved extensive experimentation with multiple AI models, focusing
on fine-tuning hyperparameters to maximize predictive accuracy. To optimize model
performance, several key techniques were employed. Grid search and Bayesian
Pag. 186 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

optimization were used to identify the most effective hyperparameters, enhancing accuracy
in energy prediction models [6]. Transfer learning leveraged pre-trained deep learning
models to improve predictive accuracy while requiring fewer training samples, a
particularly useful approach for IoT datasets with limited labeled energy records [3].
Additionally, reinforcement learning policy gradients were implemented in dynamic power
distribution systems, contributing to greater energy efficiency within smart grids [7].
Training was conducted on high-performance GPU clusters, significantly accelerating
computations and improving model convergence times.
6.4 Challenges and Future Directions in AI-Driven IoT Energy Optimization
Despite advancements in AI-driven energy management, several challenges persist. Deep
learning models require significant computational power, making real-time deployment on
low-power IoT devices difficult. The reliance on large-scale energy consumption data raises
concerns regarding privacy and potential data breaches [7]. Additionally, AI models trained
for specific IoT environments may lack scalability and struggle to adapt to different
infrastructures without retraining and fine-tuning [3]. Future research should prioritize the
development of energy-efficient AI models capable of operating on low-power devices
while maintaining high predictive accuracy.

7. Results and Performance Evaluation

Empirical implementations of AI-driven energy optimization in IoT environments have
demonstrated substantial improvements in efficiency, cost reduction, and sustainability. The
integration of machine learning (ML), deep learning (DL), and reinforcement learning (RL)
has yielded promising results in several key areas:
Energy Consumption Reduction: Studies indicate that AI-based optimization strategies
have led to an average reduction of 20-25% in energy consumption across multiple IoT
applications [5]. AI-enabled predictive analytics and real-time monitoring have contributed
to improved efficiency in power distribution and demand-response mechanisms [6].
Predictive Maintenance Accuracy: AI-powered fault detection and predictive maintenance
models have achieved over 90% accuracy, significantly reducing unplanned downtime in
industrial IoT settings. These improvements have resulted in a 30% reduction in energy
waste associated with machinery inefficiencies [3].
Grid Stability and Load Balancing: In smart grids, reinforcement learning models have
dynamically adjusted power distribution, enhancing grid stability and achieving a 15-25%
improvement in energy balancing[7]. Real-time AI algorithms have also facilitated the
seamless integration of renewable energy sources, reducing dependence on fossil fuel-based
power generation [4].
Operational Efficiency and Cost Savings: The deployment of AI-driven automation in smart
buildings has led to a 30-40% reduction in HVAC energy usage, improving climate control
Pag. 187 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

and reducing electricity costs. Studies confirm that AI-based energy scheduling in industrial
settings can optimize resource utilization, leading to a 20% decrease in operational expenses
[8].
Scalability and Adaptability: AI models trained on diverse IoT datasets have demonstrated
scalability across various applications, from residential smart meters to large-scale
industrial automation systems. Advanced federated learning techniques have enabled
decentralized AI training, minimizing data transmission overhead and enhancing real-time
adaptability [6].
The results of these studies emphasize the transformative impact of AI in energy
management for IoT systems. Future advancements in AI methodologies, including hybrid
deep reinforcement learning models and quantum AI computing, are expected to further
enhance energy efficiency and sustainability across interconnected IoT infrastructures.
Figure 3 shows a comparative bar chart illustrating energy savings across AI methodologies
and application domains

Comparative Energy Savings by AI Methodology

and Application Domain
50

40

30

20

10

0
1 2 3

ML DL (LSTM) RL Hybrid

Figure 3. Comparative Energy Savings by AI Methodology and Application Domain

ML (Machine Learning): 15–20% energy reduction (residential use)

DL (Deep Learning - LSTM): 30–35% reduction (smart buildings)
RL (Reinforcement Learning): 25% peak load reduction (smart grids)
Hybrid Models: ~35–40% average across scenarios

8. Practical Application and Solution Development

Pag. 188 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Real-world implementations have confirmed the scalability and effectiveness of AI-based

solutions, demonstrating their ability to optimize energy consumption, enhance operational
efficiency, and ensure sustainability in various IoT environments. AI-driven occupancy-
sensitive systems have transformed energy management in both commercial and residential
buildings. By employing predictive control of heating, ventilation, and air conditioning
(HVAC) along with intelligent lighting systems, energy consumption has been reduced by
30-40% [4]. Machine learning models trained on historical energy usage patterns
dynamically adjust power settings based on real-time occupancy data, significantly
improving energy efficiency [5].
In industrial IoT applications, AI-based predictive maintenance frameworks have led to
notable reductions in machine downtime and energy waste. Predictive models leveraging
deep learning-based anomaly detection algorithms have minimized energy inefficiencies in
industrial equipment by up to 25% [3]. Additionally, reinforcement learning algorithms
have been employed to optimize manufacturing processes, intelligently scheduling machine
operations to minimize energy usage while maintaining high levels of productivity [8].
The integration of AI-powered demand-response mechanisms has revolutionized modern
power grids by enabling real-time energy balancing and optimizing power distribution
across different regions. Deep reinforcement learning algorithms allow smart grids to
dynamically manage energy loads, improving balancing efficiency by 15-25% [7]. AI
models also facilitate seamless integration of renewable energy sources, reducing
dependence on fossil fuels and enhancing overall grid sustainability [6].
These applications highlight the transformative role of AI in developing energy-efficient
IoT systems, contributing to significant cost savings, reduced environmental impact, and
more resilient energy infrastructures. AI-driven technologies in smart buildings, industrial
IoT, and power grids have optimized HVAC and lighting usage, enhanced predictive
maintenance to minimize disruptions, and enabled efficient integration of renewable energy
resources, ultimately improving grid stability and sustainability.
8.1 A Novel AI-Based Framework for Automated Energy Auditing in Smart Buildings
This section proposes an advanced methodology for conducting comprehensive AI-driven
energy audits in smart buildings. The approach leverages real-time data from IoT
infrastructure, combined with artificial intelligence techniques, to deliver precise
diagnostics, automated reporting, and intelligent recommendations tailored to building-
specific energy use profiles.
8.1.1 Essential Steps in an AI-Based Energy Audit:
Audit Planning and Scoping: Define building parameters, audit objectives, expected
outcomes, and target energy efficiency KPIs.
Sensor Network Deployment: Assess existing infrastructure and install IoT sensors for
electricity, HVAC, lighting, temperature, humidity, occupancy, and CO₂ levels.

Pag. 189 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Data Aggregation and Logging: Collect real-time and historical data from all zones,
equipment types, and user activity logs.
Data Cleaning and Feature Engineering: Normalize datasets, impute missing values, and
derive custom features such as Energy Use Intensity (EUI), load curve indices, and comfort
metrics.
AI Model Integration: Use machine learning (e.g., gradient boosting, random forest) and
deep learning (e.g., LSTM, autoencoders) to predict consumption, detect anomalies, and
simulate savings scenarios.
Zone-Based Analysis: Evaluate energy performance at granular levels—by room, floor,
time segment, or equipment group—enabling highly targeted interventions.
Anomaly Detection and Diagnostics: Deploy unsupervised AI for identifying deviations
from expected operation and perform root cause analysis.
Optimization and Recommendation Generation: Provide data-driven suggestions such as
occupancy-based setpoint adjustments, predictive maintenance alerts, and automation rules.
AI-Generated Audit Report: Generate clear, stakeholder-ready reports with visual
dashboards, natural language summaries, savings projections, and ROI analyses.
Stakeholder Feedback and Iterative Looping: Incorporate user feedback into AI model
refinement and update control strategies based on validation sessions.
8.1.2 Areas of Evaluation During the Audit
The efficacy of an AI-driven energy audit is contingent upon a comprehensive and
systematic evaluation of the building's key operational and environmental subsystems. The
following domains are integral to ensuring accurate diagnostics and tailored energy
efficiency interventions:
Thermal Envelope Integrity: This includes the assessment of insulation performance,
window glazing specifications, and the extent of air infiltration, all of which influence the
building's heat retention and loss characteristics.
HVAC Operational Efficiency: Detailed analysis of heating, ventilation, and air
conditioning systems focusing on cycle frequency, energy consumption, and their
responsiveness to occupancy and environmental parameters.
Lighting System Utilization and Control: Examination of lighting patterns, alignment with
daylight availability, and the degree of automation achieved through control systems such
as timers and occupancy sensors.
Plug Load and Standby Consumption: Quantification of energy consumed by electronic
devices and appliances, with emphasis on identifying phantom loads arising from devices
left in standby mode.

Pag. 190 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Sensor and Actuator Responsiveness: Validation of the real-time accuracy and reliability of
data acquisition mechanisms, along with actuator efficiency in executing control
commands.
Renewable Energy System Performance: Evaluation of the contribution of solar, wind, or
geothermal systems, their integration with storage units, and alignment with dynamic load
demands.
Occupant Comfort and Behavioral Interaction: Analysis of user engagement with control
systems, frequency of manual overrides, and perceived thermal and lighting comfort, which
often reflect hidden inefficiencies.
8.1.3 The proposed AI-based audit framework limitations:

The proposed AI-based energy audit framework addresses the inherent limitations of
traditional audit methods by leveraging advanced technologies to deliver continuous, data-
driven performance evaluation. Through the deployment of IoT sensor networks, the system
facilitates real-time monitoring of energy flows across electricity, HVAC, lighting,
occupancy, and ambient conditions.

Machine learning algorithms are employed to generate dynamic performance benchmarks

based on historical consumption data and context-specific usage patterns, enabling accurate
assessment and comparative analysis. Furthermore, unsupervised learning techniques—
such as clustering and autoencoders—are used to detect operational anomalies and
deviations from expected performance.

The framework also automates the generation of comprehensive audit reports, providing
predictive diagnostics, optimization recommendations, and visual analytics. A key
advantage of this approach is its ability to scale granularity, allowing evaluations at the level
of individual zones, equipment, or building sections to ensure targeted energy interventions
and precise feedback.

No Audit Step AI/IoT Tools Used Expected Output

.
Pre-Audit KPI Definition, Audit
1 Planning Scope Audit Objectives & KPIs
IoT
Infrastructure Sensor Mapping, Gap
2 Assessment Analysis Sensor Deployment Plan
3 Data Collection Smart Meters, IoT Sensors Historical and Real-Time Data
Data Data Cleaning,
4 Preprocessing Normalization Cleaned & Structured Dataset
AI Model
5 Training ML/DL Algorithms Trained Models for Forecasting
EUI Calculation,
6 Zone Profiling Clustering Energy Usage Profiles
Pag. 191 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Anomaly Autoencoders, Isolation

7 Detection Forests Anomaly Flags and Reports
Recommendatio Optimization Models,
8 n Generation ROI Estimation Improvement Plan
Report
9 Compilation NLP, Visual Dashboards AI-Generated Summary

Table 2: AI-Based Energy Audit Process Overview

This framework elevates traditional energy audits into data-rich, adaptive systems that
operate autonomously, enabling continuous performance improvement. It equips facility
managers, energy analysts, and building owners with actionable insights to maximize
efficiency, comfort, and sustainability in smart building ecosystems.

9. Discussion and Future Implications

While the results of this study confirm the potential of AI in optimizing energy consumption
within IoT ecosystems, several challenges must be addressed to facilitate widespread
adoption. These include computational complexity, data privacy and security concerns,
scalability limitations, and the need for regulatory alignment.

Computational Complexity: Deep learning and reinforcement learning models offer high
performance but require significant computational resources for training and inference. This
restricts their deployment on resource-constrained IoT devices. Future research should
focus on developing energy-efficient AI models capable of operating on edge devices
without compromising accuracy. Model compression techniques such as pruning,
quantization, and knowledge distillation could play a critical role in this direction.

Data Privacy and Security: AI systems often rely on sensitive energy consumption data that
can expose user behavior patterns. Ensuring privacy-preserving AI through federated
learning and secure multi-party computation is essential. Further investigation into robust
AI models resilient to adversarial attacks is also needed.

Scalability and Interoperability: The heterogeneous nature of IoT ecosystems presents

difficulties in scaling AI models across diverse platforms. A unified framework or
middleware that supports standardized AI integration across device types and
communication protocols is essential. Research should also explore hybrid cloud-edge
architectures that balance performance and efficiency.

Policy and Regulation: As AI becomes embedded in national and industrial energy systems,
collaboration between researchers, industry stakeholders, and policymakers will be critical.
Guidelines for responsible AI deployment, transparent algorithmic decision-making, and
compliance with energy standards must be established.

Pag. 192 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Integration with Renewable Energy Systems: With the increasing penetration of solar,
wind, and other renewables, AI will play a key role in forecasting generation, optimizing
storage, and balancing supply-demand dynamics. Future systems should integrate AI-
enhanced energy forecasting and distributed energy resource management.

In summary, while AI has demonstrated strong capabilities in enhancing energy efficiency

across IoT environments, its full potential will be realized only through interdisciplinary
efforts, technological innovation, and regulatory support. The evolution of AI from isolated
models to embedded, autonomous energy managers marks a critical step toward sustainable
and intelligent energy systems.

10. Conclusion

This study demonstrated the transformative potential of artificial intelligence in optimizing

energy consumption across Internet of Things (IoT) ecosystems. Through the integration of
machine learning, deep learning, and reinforcement learning, AI models enabled predictive
analytics, dynamic control, and autonomous energy management. The proposed
methodologies were evaluated using comprehensive datasets and validated in various real-
world scenarios, including smart buildings, industrial operations, smart grids, and edge
computing environments.

Experimental results indicated that deep learning models achieved up to 35% energy
savings in building management systems, while reinforcement learning models
delivered up to 25% peak load reduction in smart grid simulations. Machine learning
approaches, although simpler, provided consistent gains in residential energy forecasting
tasks.

The findings confirm that AI-driven energy optimization not only enhances operational
efficiency but also contributes significantly to sustainability goals. The research highlighted
critical challenges related to computational demands, data security, and deployment
scalability. Future developments should focus on lightweight, privacy-preserving, and
interoperable AI frameworks that can operate efficiently in decentralized and resource-
constrained settings.

Furthermore, collaboration between academia, industry, and policymakers is vital to align

technological innovations with regulatory standards and real-world energy transition
strategies. As AI technologies mature, their integration into IoT energy infrastructures
promises to enable resilient, intelligent, and sustainable energy systems for the future.

This research contributes to the growing body of knowledge by providing a comparative

and practical evaluation of AI methodologies, identifying key implementation
considerations, and offering actionable insights for the development of next-generation
energy-aware IoT platforms.

Pag. 193 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

References
[1] Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015).
Internet of Things: A Survey on Enabling Technologies, Protocols, and Applications. IEEE
Communications Surveys & Tutorials, 17(4), 2347-2376.
[2] Mohammadi, M., Al-Fuqaha, A., Sorour, S., & Guizani, M. (2018). Deep Learning for
IoT Big Data and Streaming Analytics: A Survey. IEEE Communications Surveys &
Tutorials, 20(4), 2923-2960.
[3] F., Hussain, F., Ehatisham-Ul-Haq, M., & Azam, M. A. (2019). Machine Learning for
Energy Optimization in Smart Cities. IEEE Access, 7, 47230-47240.
[4] Zhu, Q., & Ota, K. (2021). Deep Learning-Based Energy Management in IoT
Systems. IEEE Transactions on Industrial Informatics, 17(1), 682-693.
[5] Raj, N., & Sinha, A. (2023). AI-driven Optimization for IoT-based Energy
Management. IEEE Transactions on Smart Grid, 14(3), 1525-1538.
[6] Salama, A., & Abdellatif, S. (2022). AI-driven Smart Grid Energy Management:
Challenges and Solutions. IEEE Access, 10, 17823-17838.
[7] Kushawaha, R., & Gupta, S. (2022). Reinforcement Learning for Smart Grid Energy
Optimization. Smart Grid Energy Management Journal, 18(2), 198-215.
[8] Mohammadi, M., Al-Fuqaha, A., & Sorour, S. (2018). Hybrid AI Approaches for IoT
Energy Optimization. IEEE Access, 20(4), 2923-2960.
[9] Maksimovic, M. (2017). The Role of Green Internet of Things (G-IoT) in Sustainable
Development. International Journal of Engineering and Emerging Technology, 2(2), 32–
39.

Bibliography
Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015). Internet
of Things: A Survey on Enabling Technologies, Protocols, and Applications. IEEE
Communications Surveys & Tutorials, 17(4), 2347–2376.
F., Hussain, F., Ehatisham-Ul-Haq, M., & Azam, M. A. (2019). Machine Learning for
Energy Optimization in Smart Cities. IEEE Access, 7, 47230–47240.
Kushawaha, R., & Gupta, S. (2022). Reinforcement Learning for Smart Grid Energy
Optimization. Smart Grid Energy Management Journal, 18(2), 198–215.

Pag. 194 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Maksimovic, M. (2017). The Role of Green Internet of Things (G-IoT) in Sustainable

Development. International Journal of Engineering and Emerging Technology, 2(2), 32–
39.
Mohammadi, M., Al-Fuqaha, A., & Sorour, S. (2018). Hybrid AI Approaches for IoT
Energy Optimization. IEEE Access, 20(4), 2923–2960.
Mohammadi, M., Al-Fuqaha, A., Sorour, S., & Guizani, M. (2018). Deep Learning for IoT
Big Data and Streaming Analytics: A Survey. IEEE Communications Surveys & Tutorials,
20(4), 2923–2960.
Qolomany, B., Al-Fuqaha, A., Gupta, A., Ben-Othman, J., & Qadir, J. (2019). Leveraging
Machine Learning for Sustainable Energy in Smart Cities. IEEE Communications Surveys
& Tutorials, 21(2), 1447–1484.
Raj, N., & Sinha, A. (2023). AI-driven Optimization for IoT-based Energy Management.
IEEE Transactions on Smart Grid, 14(3), 1525–1538.
Salama, A., & Abdellatif, S. (2022). AI-driven Smart Grid Energy Management: Challenges
and Solutions. IEEE Access, 10, 17823–17838.
Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. (2016). Edge Computing: Vision and
Challenges. IEEE Internet of Things Journal, 3(5), 637–646.
Xia, F., Yang, L. T., Wang, L., & Vinel, A. (2012). Internet of Things. International Journal
of Communication Systems, 25(9), 1101–1111.
Zhu, Q., & Ota, K. (2021). Deep Learning-Based Energy Management in IoT Systems.
IEEE Transactions on Industrial Informatics, 17(1), 682–693.
GSMA Intelligence. (2023). IoT Connections Forecast: 2023–2030. GSMA Publications.
International Energy Agency (IEA). (2022). Global ICT Energy Consumption Report. IEA
Digitalization & Energy Insights.
United Nations Conference on Trade and Development (UNCTAD). (2022). The
Environmental Impact of IoT Devices: Challenges and Opportunities. UNCTAD Technical
Report.

Pag. 195 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

THE DIGITAL TRAP: TEENS AND ONLINE CHALLENGES

Radu MOINESCU 1
Ciprian RĂCUCIU2
Carmen-Silvia OPRINA3

Abstract
This paper explores the alarming phenomenon of online challenges that prey on vulnerable
adolescents, particularly focusing on the notorious "Blue Whale" Challenge. It examines
how tasks desensitize participants to pain and fear, ultimately normalizing the concept of
death. Through a semantic analysis using "Tropes" software, the study highlights the
manipulative language and psychological tactics employed in these challenges. Ultimately,
the article aims to shed light on the risks associated with online challenges, advocating for
preventive measures to protect young individuals from falling into these perilous traps.
Keywords: online challenges, risk, semantic analysis, manipulation.
JEL Classification: C88, C91, D91.

1. Defining the problem

Online challenges are viral activities shared on social media and other digital platforms
where users, often teenagers, are challenged to perform dangerous, harmful or degrading
actions. These challenges include behaviors such as cyberbullying (online harassment),
catfishing (creating false identities to deceive), grooming (online manipulation for the
purpose of exploitation), as well as participating in deadly games such as "Blue Whale". In
general, online challenges take many forms, from seemingly harmless activities to extreme
acts of self-harm or anti-social behaviour.
Teenagers are more prone to online challenges due to several interconnected factors:
• Psychological factors: The emotional and cognitive development stage of
adolescence is characterized by the search for identity and the need for social proof,
which makes them more susceptible to the influence of peer groups and online
social pressures.

1PhD Candidate, National University of Science and Technology Politehnica Bucharest, Romania,
razvan_sorin.mocanu@stud.fiir.upb.ro, corresponding author
2PhD Candidate, National University of Science and Technology Politehnica Bucharest, Romania

florentina.nidelcu@stud.fiir.upb.ro
3Prof. PhD. Hab., Romanian-American University, URA, National University of Science and Technology

Politehnica Bucharest, Romania, george.carutasu@rau.ro

Pag. 196 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

• Social factors: Networks of friends and social status online play a major role in
teenagers' decisions to participate in challenges as they try to gain popularity and
appreciation in their circles.
• Technological factors: Digital platforms favor anonymity and lack of immediate
consequences, which encourages involvement in risky activities. Algorithms of
social platforms can amplify exposure to such challenges by attracting the attention
of young people.
The impact of online challenges on teenagers is significant and varied:
• Mental health: Participating in online challenges can lead to anxiety, depression,
post-traumatic stress and, in extreme cases, suicide.
• Social relationships: Teens who participate in these challenges may suffer from
social isolation, loss of trust from friends and family, and deterioration of
relationships with others.
• Academic performance: Engaging in online challenges can negatively affect
concentration and school performance, leading to absenteeism and lower school
performance.
• Physical safety: Many online challenges involve direct risks to physical health,
including self-harm, dangerous behaviors, or even death in the case of extreme
gaming.

2. Related work
The relationship between social media use and its impact on the emotional well-being of
teenagers exhibits significant gender differences. A study conducted in the UK highlights
that among teenagers aged 10 to 15, the intensive use of social networks produces varying
emotional outcomes for boys and girls. Specifically, girls tend to report a more pronounced
negative impact from digital social interactions, often experiencing increased emotional and
behavioral difficulties. In contrast, boys generally show less vulnerability to these effects.
[1]
In addition to the teenage demographic, research examining young adults aged 18 to 24
reveals that those engaging in unstructured online social activities—such as uncontrolled
use of social media—experience heightened levels of mental stress, particularly among
women. This demographic also illustrates gender disparities in social support, where young
women are significantly more affected by a lack of social support in terms of their mental
health. [2]
Popular platforms like YouTube, Instagram and TikTok have different impacts on teenagers.
YouTube, for example, is seen as having a less harmful impact compared to platforms like
Instagram, where the pressure to gain social approval through likes and comments is
greater, thus increasing social anxiety and negative comparison. Girls report more negative
emotional effects from using these platforms than boys. [3]
Cultural context plays a crucial role in shaping adolescents' experiences with social media.
A comparative study conducted in Germany, Poland, and South Africa demonstrated
Pag. 197 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

notable differences in social platform usage across these countries. For instance, Instagram
is one of the most popular platforms in South Africa, whereas Facebook remains dominant
in Germany and Poland. These cultural variances may significantly influence how
adolescents perceive and respond to online challenges, highlighting the need for a nuanced
understanding of the interplay between culture, gender, and digital interactions. [4]

3. Case study: "Blue Whale" Challenge

The "Blue Whale" Challenge (Russian: Си́ний кит) serves as a crucial case study due to its
alarming impact on teenagers worldwide and its manipulative nature. This phenomenon has
gained notoriety for its psychological manipulation, often leading participants to engage in
self-harm and other dangerous behaviors. Our study employs an in-depth semantic analysis
using "Tropes" software to identify linguistic patterns, emotions, and recurring themes
within the 50 tasks associated with the challenge. By examining these tasks, we aim to
uncover the psychological and social mechanisms that underlie online manipulation in this
context.
While existing literature has explored various aspects of the "Blue Whale" phenomenon,
our analysis distinguishes itself by concentrating on the linguistic elements of the
challenges. This approach seeks to complement current research by offering a more
nuanced understanding of how language is strategically employed to incite risky behaviors
among adolescents.
It is important to acknowledge the limitations inherent in semantic analysis; specifically, it
may not fully encapsulate the broader social and cultural contexts surrounding these
interactions. However, this method provides valuable insights into the language used to
influence vulnerable individuals.
"Tropes" is a semantic analysis and natural language processing tool designed to
automatically identify themes, concepts, and semantic relationships within text. By
analyzing language from cognitive and discursive perspectives, Tropes extracts information
about sentence structure and meaning, yielding detailed insights into the implicit messages
conveyed by the texts.
In this study, "Tropes" will be utilized to conduct a semantic analysis of the 50 tasks found
in the "Blue Whale" Challenge. The software will facilitate the identification of dominant
themes such as:
• Emotional Manipulation: Examining how language is used to evoke strong
emotional responses, potentially leading to self-harm;
• Social Isolation: Investigating language that fosters feelings of loneliness and
detachment from peers;
• Self-Esteem Pressures: Identifying phrases that create feelings of inadequacy or
compel participants to conform to harmful expectations.
Additionally, Tropes will help uncover language patterns and subliminal messages that may
exert psychological influence on teenagers, thereby elucidating the methods of coercion
embedded within the challenges.
Pag. 198 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

4. Semantic profile of "Blue Whale" tasks generated with "Tropes"

The "Blue Whale" Challenge, which emerged in Russia in 2016, attracted global attention
for its extremely dangerous nature, being associated with numerous cases of self-harm and
suicide among teenagers. It consists of a series of progressively more risky and disruptive
tasks, culminating in the final challenge of committing suicide. The challenge exploits the
psychology of the participants, giving them a sense of belonging to an exclusive group, an
aspect that is particularly appealing to vulnerable teenagers who feel socially isolated or
misunderstood.
Through semantic analysis of the tasks in this challenge, social engineering techniques were
identified that were used to mentally manipulate and control the participants. Social
engineering manifests itself by inducing feelings of loyalty and social pressure on
participants, thus forcing them to continue playing despite obvious dangers. The tasks were
designed to gradually test the limits of the participants' courage and submission, causing a
deep psychological addiction. The challenge progress was often shared publicly, creating a
sense of competition and community between participants, which further increased
engagement and motivation to complete all tasks.
This analysis highlights how online challenges such as the "Blue Whale" can use emotional
manipulation and psychological control techniques to captivate teenagers and lead them
down a dangerous path, putting their mental health and lives at risk.
An analysis of the challenge's tasks reveals the following characteristics:
• Dangerous progression: Tasks are structured to become more and more dangerous
as the challenge progresses. This progression can be seen as a method to keep the
players geared and under control, gradually increasing the level of danger and
associated risks.
• Insidiousness: Tasks are formulated in such a way that they are attractive to
vulnerable young people. Presenting self-harm as a form of emotional expression
or boundary testing can appeal to those who feel isolated or misunderstood, creating
the illusion that these actions might have a positive purpose.
• Coercion: tasks are built to control player behavior. For example, the imposition of
self-isolation may lead to an increase in the player's vulnerability to the influence
of the game, by reducing social contact and creating an increasing dependence on
the imposed tasks.
In what follows, the linguistic aspects of the tasks associated with the "Blue Whale"
phenomenon were explored in detail, with an emphasis on the persuasive and manipulative
functions of the language used. The syntactic structure, the lexical choice, as well as the
pragmatic functions of the statements were examined, in the context of a speech aimed at
inducing self-destructive behaviors.

Pag. 199 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

For this analysis, an interdisciplinary approach was used, combining tools from textual
linguistics, psycholinguistics and sociolinguistics. We analyzed each task in detail,
identifying:
• Syntactic structure: sentence types, syntactic relations, mechanisms of
subordination and coordination.
• Lexical choice: word connotations, registers of language, figures of speech.
• Pragmatic functions: speech acts (directive, expressive, declarative), implications,
assumptions.
Given the highly sensitive and controversial nature of this challenge, the semantic analysis
of the tasks required a meticulous approach. Through the "Tropes" software, we have
carried out a detailed examination of linguistic style, context, key elements, as well as the
frequency and distribution of different word classes (nouns, verbs, adjectives, pronouns,
connectors). This analysis allowed a deeper understanding of the linguistic mechanisms
used to induce and maintain the harmful behavior associated with this phenomenon.
In what follows, we will explore the linguistic features of each task in the "Blue Whale"
Challenge [5] by performing a brief semantic analysis. Attention will be directed to key
linguistic elements such as textual style, lexical register (adjectives, nouns), logical
connectors, verb moods, pronouns and verbs. The purpose of this analysis is to identify the
distinctive features of the discourse used in this challenge.
Task 1: Carve with a razor "f57" on your hand, send a photo to the curator. The text style
is imperative and straightforward, reflecting authority and urgency. The verb "carve" is
performative, indicating a harmful action, while "send" emphasizes compliance. The noun
"razor" denotes a tool for self-harm. The lack of adjectives suggests a stark and clinical
presentation of the task. The direct address "you" personalizes the command, reinforcing
accountability.
Task 2: Wake up at 4:20 a.m. and watch psychedelic and scary videos that curator sends
you. This task uses a prescriptive style that indicates routine and obligation. The adjectives
"psychedelic" and "scary" evoke emotional distress and suggest a negative experience. The
modal verb "wake up" implies a requirement for a specific time. The verbs "watch" and
"send" emphasize passive consumption and submission. The use of "curator" creates a
sense of authority and control.
Task 3: Cut your arm with a razor along your veins, but not too deep, only 3 cuts, send a
photo to the curator. The imperative tone is evident, emphasizing immediate action. The
verbs "cut" and "send" denote harmful and compliant behaviors, respectively. The phrase
"not too deep" introduces a cautionary element, revealing a disturbing normalization of self-
harm. The repetition of the noun "razor" reinforces its association with pain. The direct
pronoun "your" personalizes the act of harm.
Task 4: Draw a whale on a sheet of paper, send a photo to the curator. The task employs a
straightforward imperative style, focusing on an action with a seemingly benign outcome.
The verb "draw" suggests creativity but is juxtaposed with the task's darker context. The
noun "whale" serves as a symbolic representation within the challenge. The directive "send"

Pag. 200 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

creates an expectation of compliance, while the absence of adjectives implies a stark

presentation.
Task 5: If you are ready to "become a whale", carve "YES" on your leg. If not, cut yourself
many times (punish yourself). This task presents conditional modalities, offering two
pathways based on readiness. The verbs "carve" and "cut" highlight self-harm and
compliance, while "punish" reinforces themes of guilt. The direct pronoun "you"
personalizes the directive. The absence of descriptive adjectives creates a stark and clinical
feel, emphasizing the seriousness of the choices.
Task 6: The task with a cypher. The style here is enigmatic and cryptic, invoking curiosity
while maintaining secrecy. The noun "task" is neutral, devoid of emotional weight, while
"cypher" suggests a hidden meaning. The lack of verbs indicates passivity and mystery. The
absence of explicit adjectives or pronouns creates an impersonal and abstract directive.
Task 7: Carve "f40" on your hand, send a photo to the curator. This task employs a direct
imperative style, emphasizing compliance. The verb "carve" connotes self-harm, while the
lack of adjectives suggests a stark presentation. The use of the noun "curator" implies a
hierarchical structure. The direct command reinforces personal accountability through the
use of "your."
Task 8: Type "#i_am_whale" in your VKontakte status. The directive is straightforward,
using an imperative tone. The verb "type" denotes a digital action, highlighting engagement
with social media. The hashtag serves as a symbol of identity, linking the participant to the
challenge. The absence of adjectives contributes to a clinical presentation, while the
pronoun "your" personalizes the task.
Task 9: You have to overcome your fear. The task employs a prescriptive style that
emphasizes personal growth and challenge. The verb "overcome" is action-oriented and
empowering, though it reflects pressure to conform. The noun "fear" embodies emotional
complexity, while the use of "you" personalizes the directive, fostering a sense of individual
responsibility.
Task 10: Wake up at 4:20 a.m. and go to a roof (the higher the better). The task's imperative
style emphasizes routine and urgency. The verb "go" signifies movement and risk, while
the phrase "the higher the better" introduces a competitive element. The nouns "roof" and
"higher" evoke danger and elevation. The use of "you" personalizes the instruction,
reinforcing accountability.
Task 11: Carve a whale on your hand with a razor, send a photo to the curator. This task
utilizes a stark imperative tone, emphasizing self-harm through the verb "carve." The noun
"whale" symbolizes identity within the challenge, while "razor" signifies danger. The
command to "send" reinforces compliance. The absence of adjectives creates a clinical
atmosphere.
Task 12: Watch psychedelic and horror videos all day. The task employs a prescriptive style
that suggests continuous exposure to disturbing content. The adjectives "psychedelic" and
"horror" evoke fear and confusion. The verb "watch" implies passive engagement. The

Pag. 201 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

absence of pronouns personalizes the command while creating distance from the emotional
impact.
Task 13: Listen to music that “they” (curators) send you. The task is structured as an
imperative, emphasizing compliance. The verb "listen" denotes passive engagement, while
the pronoun "they" creates an ambiguous authority figure. The absence of adjectives
suggests a stark, emotionless directive. The use of "you" would enhance personal
responsibility but is absent here.
Task 14: Cut your lip. This imperative command uses a blunt style, emphasizing self-harm.
The verb "cut" denotes a harmful action, while the noun "lip" personalizes the area of harm.
The absence of adjectives and connectors suggests a clinical presentation, reinforcing the
gravity of the task.
Task 15: Poke your hand with a needle many times. The imperative tone is clear, focusing
on painful actions. The verb "poke" suggests a repetitive and harmful act, while "needle"
evokes imagery of danger. The absence of adjectives emphasizes the starkness of the task.
The lack of personal pronouns creates a distance from the emotional consequences.
Task 16: Do something painful to yourself, make yourself sick. This task presents an
ambiguous imperative style, focusing on self-inflicted harm. The verbs "do" and "make"
are vague yet emphasize a call to action. The absence of explicit adjectives renders the task
stark, while the use of "yourself" personalizes the directive, inviting introspection into one's
actions.
Task 17: Go to the highest roof you can find, stand on the edge for some time. The
imperative tone underscores risk-taking behavior. The verb "go" indicates movement
toward danger, while "stand" implies vulnerability. The phrase "highest roof" evokes height
and peril. The lack of adjectives and explicit connectors suggests a stark, direct command.
Task 18: Go to a bridge, stand on the edge. This task uses a direct imperative style,
emphasizing risk. The verbs "go" and "stand" indicate physical actions associated with
danger. The noun "bridge" evokes imagery of transition and risk. The absence of adjectives
heightens the starkness of the directive.
Task 19: Climb up a crane or at least try to do it. The task employs a prescriptive style,
focusing on physical challenge and risk. The verb "climb" indicates a high-stakes action,
while "try" introduces an element of uncertainty. The noun "crane" symbolizes elevation
and danger. The lack of adjectives renders the command stark and direct.
Task 20: The curator checks if you are trustworthy. This task presents a declarative
structure, emphasizing evaluation. The verb "checks" implies surveillance and control,
while "trustworthy" introduces a moral dimension. The pronoun "you" personalizes the
directive, reinforcing individual accountability. The absence of adjectives contributes to a
neutral tone.
Task 21: Have a talk “with a whale” (with another player like you or with a curator) in
Skype. The directive employs an imperative tone focused on communication. The verbs
"have" and "talk" indicate social interaction, while the phrase "with a whale" symbolizes

Pag. 202 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

peer connection within the challenge. The use of parentheses adds an explanatory note, and
the lack of adjectives creates a neutral tone.
Task 22: Go to a roof and sit on the edge with your legs dangling. The imperative style
emphasizes risky action. The verbs "go" and "sit" imply physical movement and
vulnerability. The phrase "with your legs dangling" introduces imagery of danger and
instability. The absence of adjectives reinforces the starkness of the command.
Task 23: Another task with a cypher. This task employs an ambiguous and cryptic tone. The
noun "task" remains neutral, while "cypher" suggests secrecy and hidden meanings. The
absence of verbs and explicit adjectives creates a sense of mystery, emphasizing the
enigmatic nature of the directive.
Task 24: Secret task. The use of "secret" implies a hidden directive that is both enticing and
ominous. The noun "task" is neutral, devoid of emotional weight. The absence of verbs and
adjectives contributes to an air of mystery and suspense, suggesting that the nature of the
task remains concealed, thereby fostering intrigue while simultaneously reinforcing the
isolation felt by participants.
Task 25: Have a meeting with a “"whale". This task adopts an imperative structure focused
on social interaction. The verb "have" indicates a directive for engagement, while "meeting"
connotes a formal gathering. The use of the term “whale” symbolizes a peer within the
challenge, creating a sense of belonging. The lack of adjectives and connectors maintains a
straightforward tone, emphasizing compliance.
Task 26: The curator tells you the date of your death and you have to accept it. The task
employs a chilling imperative style, focusing on acceptance of mortality. The verb "tells"
indicates a transfer of information, while "accept" introduces an emotional dimension,
urging compliance. The phrase "date of your death" carries heavy existential weight. The
use of "you" personalizes the command, reinforcing individual accountability in the face of
a terrifying directive.
Task 27: Wake up at 4:20 a.m. and go to rails (visit any railroad that you can find). The
directive uses an imperative tone that emphasizes a specific time and place for risky
behavior. The verbs "wake up" and "go" indicate required actions, while "rails" evokes
imagery of danger and potential harm. The absence of adjectives lends a starkness to the
task, reinforcing its immediacy and seriousness.
Task 28: Don’t talk to anyone all day. This task employs a straightforward imperative style,
emphasizing isolation. The verb "talk" indicates communication, while "anyone" highlights
a complete withdrawal from social interactions. The lack of adjectives suggests a clinical
presentation of loneliness, reinforcing the emotional impact of disconnecting.
Task 29: Make a vow that "you’re a whale". The task presents a prescriptive structure,
emphasizing commitment. The verb "make" indicates an active choice, while "vow" implies
a solemn promise. The phrase "you’re a whale" symbolizes identification with the
challenge, creating a sense of belonging. The absence of adjectives reinforces a stark
commitment without emotional embellishment.

Pag. 203 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Task 30-49: Every day you wake up at 4:20 a.m., watch horror videos, listen to music that
"they" send you, make 1 cut on your body per day, talk "to a whale". These tasks utilize a
repetitive and ritualistic structure, emphasizing the monotony of compliance. The verbs
"wake up", "watch", "listen", "make", and "talk" denote a routine of harmful actions and
passive consumption. The use of the pronoun "you" personalizes the experience, making
the participant feel implicated. The lack of adjectives contributes to a clinical atmosphere,
normalizing self-harm within the context of the challenge. The phrase "4:20 a.m."
establishes a specific, eerie ritual that heightens the sense of obligation.
Task 50: Jump off a high building. Take your life. This task employs a stark and chilling
imperative tone, signaling a culmination of escalating violence. The verbs "jump" and
"take" indicate definitive actions that lead to irreversible consequences. The noun
"building" evokes imagery of height and danger, while "life" underscores the existential
weight of the decision. The absence of adjectives strips away emotional nuance, rendering
the task a blunt command devoid of empathy. The lack of pronouns creates a sense of
detachment, reinforcing the chilling nature of the directive and the totalizing power of the
challenge.

5. Key findings
The semantic analysis conducted using "Tropes" software provides a comprehensive
examination of the language, structure, and psychological implications embedded within
the 50 tasks of the "Blue Whale" Challenge. The following findings synthesize the analysis
into key categories: text style, contextual placement, representative elements, and the main
categories of words, including nouns, verbs, adjectives, pronouns, and connectors.
The text style of the tasks is characterized by prescriptive and imperative language, which
conveys authority and urgency. The starkness of the directives, often devoid of emotional
qualifiers, reinforces a sense of desensitization to violence and self-harm. This stylistic
choice cultivates an atmosphere of coercion, as the language serves to normalize harmful
behaviors within a structured framework. The repetition of commands across tasks further
amplifies their weight, framing them as obligatory rather than optional actions.

Fig. 1. Text style

Pag. 204 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Contextually, the tasks are situated within a narrative arc that progressively escalates in
intensity and severity, culminating in directives that advocate for extreme self-harm and
suicidal ideation (Fig. 1). This escalation serves to desensitize participants to the emotional
and psychological consequences of their actions. The temporal placement of certain tasks,
such as those occurring at early morning hours, suggests an element of ritualism,
encouraging participants to embrace discomfort and isolation as a means of fulfillment
within the challenge. This context heightens the psychological burden on individuals,
fostering feelings of loneliness and despair.
The representative elements within the tasks are emblematic of themes surrounding
mortality, self-identity, and existential struggle. Symbols such as the "whale" serve to
encapsulate notions of vastness and depth, while tasks involving horror media, self-harm,
and social withdrawal illustrate a disturbing convergence of fear, isolation, and community
engagement. The symbolic language invites introspection regarding personal identity and
societal implications, thereby deepening the emotional resonance of the directives.
The nouns employed throughout the tasks—such as "whale," "building," "life," and
"curator"—carry significant connotations that reflect the overarching themes of the
challenge. The use of concrete nouns establishes a direct relationship with the actions
mandated by the tasks, emphasizing the existential weight of the decisions presented. The
broad scope of social nouns highlights the pervasive nature of isolation inherent in the
challenge.
The verb categories (Fig. 2) indicate a predominance of factive verbs (59.6%), suggesting
a focus on actions that denote established truths or realities. This is contrasted by a smaller
percentage of stative verbs (6.7%), which imply a condition or state, and reflexive verbs
(33.7%), indicating actions directed back at the subject. The absence of performative verbs
(0%) emphasizes the lack of agency granted to participants, who are compelled to act rather
than to express or create through their actions.

Fig. 2. The relationship of verbs

Pag. 205 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The use of adjectives is minimal (Fig. 3), with objective adjectives (5.3%) and subjective
adjectives (5.3%) present in only a small fraction of the tasks. This scarcity contributes to
the starkness of the language, allowing the emotional gravity of the tasks to resonate without
the influence of descriptive embellishment.

Fig. 3. The relationship of adjectives

The use of pronouns predominantly includes second-person references (Fig. 4), which
establish direct engagement with the participant and foster personal accountability. The
presence of conditional connectors (16.7%) suggests that some tasks are framed as
contingent on certain behaviors or outcomes, although the absence of causal connectors
(0%) indicates a lack of explanatory context. The occurrence of disjunctive connectors
(16.7%) reflects the dichotomies presented within the tasks, underscoring the conflicting
emotions and choices faced by participants.

Fig. 4. The relationship of pronouns

Pag. 206 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

6. Conclusion
The proliferation of online challenges, particularly those that promote self-harm and risky
behaviors, poses a significant threat to the mental health and well-being of adolescents. As
evidenced by the analysis of the "Blue Whale" Challenge and similar phenomena, these
digital traps often exploit vulnerabilities associated with youth, such as the search for
identity, peer acceptance, and a desire for belonging. The stark language and coercive
structures inherent in these challenges reflect a concerning trend in digital spaces that can
lead to devastating consequences.
To mitigate the risks associated with online challenges, it is imperative to implement
comprehensive preventive measures aimed at protecting young people from falling into
these dangerous traps.
Developing educational programs that focus on digital literacy is essential. These programs
should educate teens about the nature of online challenges, emphasizing critical thinking
skills to recognize manipulative content and the potential dangers of engaging with such
activities. Workshops in schools can provide a safe space for discussions about peer
pressure, mental health, and the risks of online participation.
Encouraging active involvement from parents and guardians is crucial. Parents should be
educated on the digital landscape their children navigate, including awareness of popular
online challenges and the signs of distress in their teens. Open communication between
parents and teens regarding online experiences can foster an environment where young
people feel safe discussing their encounters with potentially harmful content.
Creating peer support networks can empower young people to resist the allure of dangerous
challenges. Schools and community organizations can facilitate programs that encourage
teens to speak out against harmful activities, fostering a culture of mutual support and
accountability. By normalizing conversations around mental health and online safety, these
networks can help teens feel less isolated and more connected.
Increased access to mental health resources is critical. Schools should provide resources
such as counseling services and helplines for students facing emotional distress. Educating
teens about the availability of these resources can encourage them to seek help if they are
struggling with feelings of loneliness or depression, which are often exploited by online
challenges.
Collaboration between mental health organizations, educators, and technology companies
can lead to the development of tools that promote positive online interactions. These tools
could include features that allow users to report harmful content easily, as well as
educational prompts that encourage users to reflect on their online engagement before
participating in challenges.
As the digital landscape continues to evolve, so too must our strategies for safeguarding
young people. By combining education, parental involvement, supportive peer networks,
and accessible mental health resources, we can create a multifaceted approach to prevent
teens from falling into the traps of online challenges. Ultimately, fostering a culture of
Pag. 207 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

resilience, awareness, and open communication will empower young people to navigate the
digital world safely, promoting their overall well-being and development in the process.

References
[1] Cara L. BOOKER, Yvonne J. KELLY, Amanda SACKER – Gender differences in the
associations between age trends of social media interaction and well-being among 10-15
year olds in the UK, BMC Public Health, Volume 18, Article number 321, 20 March
2018, DOI: 10.1186/s12889-018-5220-4, ISSN: 1471-2458, web resource:
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-018-5220-4, accessed
on 28 September 2024
[2] Rune JOHANSEN, Mari Nicholls ESPETVEDT, Heidi LYSHOL, Jocelyne
CLENCH-AAS, Ingri MYKLESTAD – Mental distress among young adults – gender
differences in the role of social support, BMC Public Health, Volume 21, Article number
2152, 24 November 2021, ISSN: 1471-2458, DOI: 10.1186/s12889-021-12109-5, web
resource: https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-021-12109-
5, accessed on 28 September 2024
[3] Emily A. VOGELS, Risa GELLES-WATNICK – Teens and social media: Key
findings from Pew Research Center surveys, Pew Research Center, 24 April 2023, web
resource: https://www.pewresearch.org/short-reads/2023/04/24/teens-and-social-media-
key-findings-from-pew-research-center-surveys/, accessed on 28 September 2024
[4] Kaja J. FIETKIEWICZ, Elmar LINS, Adheesh BUDREE – Investigating the
Generation- and Gender-Dependent Differences in Social Media Use: A Cross-Cultural
Study in Germany, Poland and South Africa, Social Computing and Social Media.
Technologies and Analytics (SCSM 2018), Publisher Springer, Cham, 31 May 2018, pp.
183–200, DOI: 10.1007/978-3-319-91485-5_14, Print ISBN: 978-3-319-91484-8, web
resource: https://link.springer.com/chapter/10.1007/978-3-319-91485-5_14, accessed on
28 September 2024
[5] Jonathan TAYLOR – What is the Blue Whale Challenge?, Educare.co.uk, web
resource:
https://www.educare.co.uk/files_cms/resources/What_is_the_blue_whale_challenge.pdf,
accessed on 28 September 2024

Bibliography
Cara L. BOOKER, Yvonne J. KELLY, Amanda SACKER – Gender differences in the
associations between age trends of social media interaction and well-being among 10-15
year olds in the UK, BMC Public Health, Volume 18, Article number 321, 20 March
2018, DOI: 10.1186/s12889-018-5220-4, ISSN: 1471-2458, web resource:
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-018-5220-4, accessed
on 28 September, 2024
Rune JOHANSEN, Mari Nicholls ESPETVEDT, Heidi LYSHOL, Jocelyne CLENCH-
AAS, Ingri MYKLESTAD – Mental distress among young adults – gender differences in
Pag. 208 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

the role of social support, BMC Public Health, Volume 21, Article number 2152, 24
November 2021, DOI: 10.1186/s12889-021-12109-5, ISSN: 1471-2458, web resource:
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-021-12109-5,
accessed on 28 September, 2024
Emily A. VOGELS, Risa GELLES-WATNICK – Teens and social media: Key findings
from Pew Research Center surveys, Pew Research Center, 24 April 2023, web resource:
https://www.pewresearch.org/short-reads/2023/04/24/teens-and-social-media-key-
findings-from-pew-research-center-surveys/, accessed on 28 September, 2024
Kaja J. FIETKIEWICZ, Elmar LINS, Adheesh BUDREE – Investigating the Generation-
and Gender-Dependent Differences in Social Media Use: A Cross-Cultural Study in
Germany, Poland and South Africa, Social Computing and Social Media. Technologies
and Analytics (SCSM 2018), Publisher Springer, Cham, 31 May 2018, pp. 183–200, DOI:
10.1007/978-3-319-91485-5_14, Print ISBN: 978-3-319-91484-8, web resource:
https://link.springer.com/chapter/10.1007/978-3-319-91485-5_14, accessed on 28
September, 2024
Jonathan TAYLOR – What is the Blue Whale Challenge?, Educare.co.uk, web resource:
https://www.educare.co.uk/files_cms/resources/What_is_the_blue_whale_challenge.pdf,
accessed on 28 September, 2024
Anshu MALHOTRA, Rajni JINDAL – Multimodal Deep Learning Architecture for
Identifying Victims of Online Death Games, Published in Data Analytics and
Management, Proceedings of ICDAM, Publisher Springer, 26 June 2021, pp. 827–841,
DOI: 10.1007/978-981-15-8335-3_63, Print ISBN: 978-981-15-8334-6, Online ISBN:
978-981-15-8335-3, web resource:
https://www.researchgate.net/publication/348209895_Multimodal_Deep_Learning_Archit
ecture_for_Identifying_Victims_of_Online_Death_Games, accessed on 28 September,
2024
Richa MUKHRA, Neha BARYAH, Kewal KRISHAN, Tanuj KANCHAN – ‘Blue Whale
Challenge’: A Game or Crime?, Science and Engineering Ethics, Publisher Springer, 11
November 2017, Volume 25, pp. 285–291, DOI: 10.1007/s11948-017-0004-2, web
resource:
https://www.researchgate.net/publication/321014625_'Blue_Whale_Challenge'_A_Game_
or_Crime, accessed on 07 august 2024
Anirban MUKHERJEE, Satish SHUKLA – Decoding the Blue Whale Challenge from
Sociological Perspectives, Journal of National Development, Volume 31, No. 1, 1 July
2018, pp. 112-119, DOI: 10.29070/31/57444, web resource:
https://www.researchgate.net/publication/331369208_Decoding_the_Blue_Whale_Challe
nge_from_Sociological_Perspectives, accessed on 07 august 2024
Pierre MOLETTE – De l’APD à Tropes: comment un outil d’analyse de contenu peut
évoluer en logiciel de classification sémantique généraliste, Communication au colloque
Psychologie et communication Tarbes, June 2009, web resource:
https://www.tropes.fr/PierreMolette-CommunicationColloquePsychoTarbesJuin2009.pdf,
accessed on 07 august 2024
Pag. 209 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Catherine KERBRAT-ORECCHIONI – L'énonciation: De la subjectivité dans le langage,

Publisher Arman Colin, 2009, ISBN: 9782200243913
Patrick CHARAUDEAU – Grammaire du sens et de l'expression, Publisher Hachette-
Education, 1992, ISBN: 978-2-35935-241-2
Kurt LEWIN – Frontiers in Group Dynamics: Concept, Method and Reality in Social
Science; Social Equilibria and Social Change, Human Relations, Volume 1, Issue 1, 1
June 1947, Sage Publications Ltd., DOI: 10.1177/001872674700100103, pp. 5–41, web
resource: https://lchc.ucsd.edu/MCA/Mail/xmcamail.2013_07.dir/pdfeF83xvxgaM.pdf,
accessed on 07 August 2024
Robert-Vincent JOULE, Jean-Leon BEAUVOIS – Mic tratat de manipulare pentru uzul
oamenilor cinstiți, 2016, Editura Polirom, ISBN: 978-973-46-6173-2

Pag. 210 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

DRAWING ON AUTOCAD THE HORN OF THE ICONIC ALTEC LANSING

LOUDSPEAKER VOICE OF THE THEATER A7

Jose MUJICA1

Abstract
The horn Altec Lansing model H-811B is one of the main components of the loudspeakers
A-7 and Model 19 of this milestone brand. It is not usual that lovers and collectors of a
brand of loudspeakers have Internet sites with over 100K followers 44 years after stopping
the manufacturing of these classic models. Nowadays, we find that some followers have
opened factories to build them anew. However these facilities have made the low frequency
woofers and mid-high frequency drivers, but not the horns and the boxes. About the boxes,
the blueprints and its construction do not represent a major challenge, after all, in more than
75 years of the first model of the Altec Lansing series Voice Of The Theater, a legion of
follower have learned how to make the wood curves as the original. But the cast aluminum
horns are a different matter because of the difficulties involved in their process. Some of
the followers have built them using wood and Fiberglass horns but very few using
aluminum. Neither of the blueprints found have information about the curves of the horn,
only the lineal dimensions. For this reason and knowing that there is not a 3D AutoCad
drawing on the Internet of this horn, at least one that I could find, I decided to write a
method that can be helpful to draw a prototype model that could help enthusiasts and
collectors to build the horns using aluminum.
Keywords: Autocad, Computer-Aided Design, Computer Drawing Method, Command
Bugs in Design Software.
JEL Classification: C61

Introduction
For this paper the author used the 1968 blueprints of the Altec Lansing loudspeaker sectoral
horn, model 811B [1][2][3]. The blueprints that I found only show the lineal dimensions
of the horns. I was not able to find any information about the different curvatures of the
horns. (See Figure 1).

1PhD Candidate, National University of Science and Technology Politehnica Bucharest, Romania,
razvan_sorin.mocanu@stud.fiir.upb.ro, corresponding author
Pag. 211 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 3 Dimension of the Sectoral horn 811B

I also used a real Altec Lansing horn, belonging to Model 19 of this brand, in order to get
better accuracy for the goal of this paper. Any stage of the drawing was compared with the
real horn shown in Figure 2.

Figure 4 Real Altec Lansing 811B sectoral horn

Curves are ubiquitous in nature, design, and fields such as mathematics. In general, drawing
curves is a challenging task. For this reason, the main goal of this work is focused on the
reproduction of the main body of the horn. In all drawings I have used units that attempt to
preserve the external dimensions of the horn.

1. Drawing a reproduction of the horn

1.1 Drawing the side curves of the horn

To draw the horn side curve, the author used the command ARC with three points of
AutoCad [4]. To get the coordinates of the three points, I initially thought of applying the
Least Square Method to obtain the curve equation, but I estimated that the precision would
be the same as graphing these three points over the Cartesian plane. See Figure 3.

Pag. 212 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 5 Horn top curve of the 811B horn

The coordinates of three points obtained were (0,0.75), (3.65,1.3), (7.3,3.5). Once the upper
curve was obtained I proceeded to use the MIRROR command to generate the lower curve.
Then using the MOVE command I displaced the lower curve 1.5 units down along the Y
axis. See Figure 4.

Figure 6 Using the MIRROR and MOVE commands to locate the upper and lower curve
on the plane.

1.2 Extruding the up and down side curves.

To extrude the curves in 3D we should apply the 3DFACE command to join both curves to
create a face. Having the face, we use the EXTRUDE command 15.5 units along the Z axis
to get the wider view of the horn, that will be its mouth. As seen in Figure 5.

Pag. 213 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 7 Extrude the side face

1.3 Cutting side planes.

Now we generate two planes, each one perpendicular to the XZ plane. Both inclined with
respect to the Y axes.
These planes will be the side walls of the horn. To get their coordinates I used the only
blueprint that I could find on the Internet with a top view of the model 811B. [2]. See Figure
6.

Figure 8 Top view of the 811B horn.

The coordinates of the first plane are (0.3.15,-7), (7.3,3.15,0), (7.3,-3.5,0) and (0,-3.5,-7).
The coordinates of the second plane are (0, 3.5,-8.5), (7.3, 3.5,-15.5), (7.3,-3.15,-15.5) and
(0,-3.5,15.5). Both boundaries were made using the 3DPOLY command. After the
command 3DFACE was used to create the planes. See Figure 7.

Pag. 214 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 9 3D view of the two planes.

Once obtained the two planes across the solid I proceeded to use the CONVTOSURFACE
command so that all the components were surfaces. See Figure 8.

Figure 10 Inclined base of the horn walls

To proceed to remove the leftover parts of the body of the horn I used the command
SLICE and its option Surface. See Figure 9.

Figure 11 Getting the angle of the inclined panel

Pag. 215 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

1.4 Top and bottom of the horn.

The top and the bottom of the horn are not flat, they are curved, so we need to get rid of
the actual up and low surfaces and draw the suitable ones.
To create the top and bottom flat curved surfaces, the first step is to create two flat curves
on the mouth using the ARC command with three points. The coordinates of the upper
curve are (7.3, 3.5,0), (7.3,1.25,-7.75) and (7.3,3.5,-15.5). The coordinates of the lower
curve are (7.3, -3.5, 0), (7.3, -1.25, 7.75) and (7.3,-3.5,-15.5). Because the ARC command
is two dimensional, I only used the Y and Z coordinates. (See figure 10).

Figure 12 Curve walls sliced

Once the two flat curves of the mouth have been obtained I deleted the actual flat up and
low surfaces. Now I need to create a curve region which boundaries are the flat curve of the
mouth that I drew before, the two inclined side walls and a line that I draw on the back of
the horn as shown in Figure 11.

Figure 13 Line at the end of the horn to close the region

Pag. 216 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The command EDGESURF that creates a mesh (A 3D surface) between four contiguous
edges or curves that form a close loop, was used to generate the region. Before using this
command, I need to get the edges from the side walls to generate the four sides needed to
close the region. For this purpose, I use the command XEDGES, (Extract Edges in the 3D
Tools Menu/Solid Editing.). I perceived also that the command EXPLODE works well to
break down a compound object into its individual components.

1.5 Drawing the brims of the mouth.

To create the extension brim of the mouth of the horn, I must draw a line between the front
limits of the mouth and then use the command REGION (Menu Draw.). Selecting this last
line and the flat curve that I created in the section 2.4 I generated a front surface. (See Figure
12).

Figure 14 Front surface color in red to create the brim.

To finish the brim I rotate the front surface 90° using the command REVOLVE around the
axis formed by the line that joins the side boundaries as shown in Figure 13.

Figure 15 The first brim showed

Pag. 217 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

I can draw the second brim and make visible all the sides of the horn as show in Figures
14 and 15.

Figure 16 View of the first side wall

Figure 17 View of both panels with realistic visual style

1.6 Drawing the vanes of the horn.

The vanes dimension and angles were measured from my own 811B horn, showed in Figure
2. Because the scope of this work is the methodology to draw the body of the horn, I did
not build any special device or tool to get the correct horn curves measurements.

With the measurements shown in Figure 16 the angle of the incline vane is 30°.

Pag. 218 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 18 Angle of the inclined vanes

The vanes are symmetrically distributed, the center vane coordinates are (7.3,3.5,7.75),
(9.55,3.5,7.75), (9.55,-3.5,7.75) and (7.3,-3.5,7.75) as showed in Figure 17. To draw the
plane of the vane I used the 3DFACE command.

Figure 19 Coordinates of the center vane

With the generated planes it was necessary to cut off the spare over the brim using the
SLIDE command as I did at the end of 2.3 section.
The coordinates of the inclined left and right vanes are as follows: Left one ( 7.3,
3.5,3.5.174), (9.55, 3.5, 3.875), (9.55,-3.5,3.875) and (7.3,-3.5,5.174). Right one (7.3,
3.5,10.3260), (9.55, 3.5, 11.6250), (9.55,-3.5, 11.6250) and (7.3, -3.5, 10.3260). Once the
inclined vanes were drawn and the spare parts were cut off I used the SLICE command to
obtain the view of Figure 18.

Pag. 219 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 20 Body of the horn without spare parts

2. Drawing the throat

Now proceed to draw the throat with a depth of 2.95 units. I started from the end of the side
curves previously obtained. The coordinates to get the additional points to create the four
3DFaces needed to build the throat are, (-2.95,0.625,-7.125), (-2.95, 0.625, -8.375), (-2.95,-
0.625,-8.375) and (-2.95,-625,-7.125), as showed in the Figure 19.

Figure 21 Building the throat

2.1 Drawing the base of the driver.

The base of the driver is a circle with the center on the axis of the horn at (-2.95,0,-7.75).
Once created , I used 3DROTATE to make it perpendicular to the axis of the horn. See
Figure 20.

Pag. 220 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 22 Circle as Driver base to rotate

Because the base of the driver is perpendicular to the axis of the horn, I used the EXTRUDE
command with 0.2 depth and converted the object to solid with the CONVTOSOLID
command. See Figure 21.

Figure 23Circular solid base

I have to repeat all the previous procedures to create a smaller circular hole at the center of
the solid base. I used the CIRCULAR command with the Radius and Center coordinate
options. The Radius is 0.5 units, and the center is at (-2.95,0,-7.75). After the EXTRUDE
command I used the SUBTRACT command to show the empty space. Alternatively, I could
have used the PRESSPULL command also to get the hollow, but this command is
cumbersome to use because one has to follow one particular sequence of instructions to get
the desired effect. See Figure 22.

Pag. 221 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 24 Driver base hollow

3. Drawing the driver

Because the main purpose of this work is to show a methodology to draw the curves of the
body of the horn, I only wanted to draw the driver as a non-detailed visual reference in the
final drawing. For this purpose, I only considered the contour side section view of an Altec
Lansing driver specification sheet[5]. See Figure 23.

Figure 25 Specification sheet of Altec 807 and 808 drivers

From the shape of the image of Figure 23, I drew an approximation of a half of the driver
using AutoCad. See Figure 24.

Figure 26 Approximation of a half of the Altec driver

Pag. 222 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Now I use the REVOLVE command to get the driver in 3D. See Figure 25.

Figure 27 Driver sectional view

I created the driver scaled in a single file and used copy and paste to join with the figure in
the horn file. At this step I must use the 3DROTATE and ALIGN commands to connect the
driver to the horn.

4 End view of the entire horn and driver

Finally, I have the entire drawing of the horn with its driver as shown in the following
Figures.

Figure 28 Full wired view of the horn

Pag. 223 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 29 Full conceptual view of the horn

5 Conclusions
The view of the Altec Lansing horn 811B never cease to astonish its followers. It not only
reaches the collector micro universe but also the general audiences who admire the unique
characteristics of this marvelous horn.
New technologies in the midrange horns surpasses the sound quality of the Altec Lansing
horn 811B but not the experience to see and hear it. You can compare this with the
experience of seeing and listening to a Gramophone or a vintage vinyl record player.
Therefore, Altec Lansing horn 811B followers will continue growing through the years to
come. It is my hope that this paper can be used and be useful as a reference or starting point
for those who in the future will attempt to get the perfect dimensions of the Altec Lansing
horn 811B.

Acknowledgement
The author wishes to express his gratitude to Dr. Ramon Mata-Toledo (James Madison
University) for his thorough review and constructive criticism of this work. His input,
highly appreciated by the author, contributed to improve this work significantly.

References
[1] Altec Lansing, Loudspeaker Enclosures. Their Design and Use. Altec lansing AL-
6030-1.
https://www.lansingheritage.org/images/altec/plans/1974-enc-design/page20.jpg
[2] Altec Lansing, Loudspeaker Enclosures. Their Design and Use. CP-572-10K. LTV
LING. 1968
https://www.lansingheritage.org/images/altec/plans/1968-enclosure/page07.jpg
[3] Altec Lansing, Loudspeaker Enclosures, their use and design, CAT NO AL 1307-2.
https://www.lansingheritage.org/images/altec/plans/1960's-lf-design/page10.jpg

Pag. 224 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[4] Randy H. Shih, AutoCAD 2021 Tutorial Second Level 3D Modeling, SDC
Publications, 2020
[5] Altec Lansing, 1968 Enclosure Plans, Audio Heritage, CP 1070 -5K, AL-1777-1. 1968
https://www.lansingheritage.org/images/altec/specs/components/807-808/page1.jpg

Bibliography

DAVIS DON, EUGENE PATRONICS, -Sound System Engineering, ISBN-13:978-0-

240-8083-7, Elsevier, 2006.
EARGLE JOHN, FOREMAN CHRIS, -Audio Engineering For Sound Reinforcement,
ISBN 0-634-04355-2, Hal Leonard Corporation, 2002.
HASS JOEL, HELL CHRISTOPHER, PRZEMYSLAW BOGACKI, MAURICE WEIR,
GEORGE B. THOMAS, -University Calculus Early Transcendentals, ISBN: 13:978-0-
13-516487-7, Pearson, 2020.
SHIH H. RANDY, AutoCAD 2021 Tutorial Second Level 3D Modeling, ISBN: 978-1-
63057-357-7, SDC Publications, August 18, 2020.
TREMAINE HOWARD, Audiocyclopedia, ISBN: 0-672-20675-7, Howards W. Sams &
Co, Inc., 1979.
ZILL DENNIS, Advanced Engineering Mathematics, ISBN: 9781284206241, Jones
&Bartlett Learning, 2022.
ALTEC LANSING, Enclosure Plans, Audio Heritage, CP 1070 -5K, AL-1777-1. 1968
https://www.lansingheritage.org/images/altec/specs/components/807-808/page1.jpg
ALTEC LANSING, Loudspeaker Enclosures. Their Design and Use. Altec lansing AL-
6030-1. https://www.lansingheritage.org/images/altec/plans/1974-enc-design/page20.jpg
ALTEC LANSING, Loudspeaker Enclosures. Their Design and Use. CP-572-10K. LTV
LING. 1968https://www.lansingheritage.org/images/altec/plans/1968-
enclosure/page07.jpg
ALTEC LANSING, Loudspeaker Enclosures, their use and design, CAT NO AL 1307-2.
https://www.lansingheritage.org/images/altec/plans/1960's-lf-design/page10.jpg

Pag. 225 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

GENERATIVE PRETRAINED TRANSFORMERS FOR INVESTOR-CENTRIC

PORTFOLIO CONSTRUCTION

Dimitrios Papakyriakopoulos1
Manolis Kritikos2

Abstract
The objective of this paper is to examine the application of generative artificial intelligence
in personalized portfolio construction and evaluate its performance relative to traditional
benchmarks. A generative AI model, specifically OpenAI’s GPT-4o, was employed to
construct investment portfolios for ten virtual investor profiles over a fixed three-month
investment horizon. The methodology involved prompting the model to create portfolio
allocations, followed by performance evaluation using financial metrics including total
return, volatility, beta, Sharpe ratio, and maximum drawdown. All AI-generated portfolios
outperformed the S&P 500 index over the investment period, demonstrating stronger risk-
adjusted returns and lower drawdowns. These results highlight the potential of large
language models to synthesize financial data and produce competitive investment
strategies. The study contributes to the growing body of research on AI-driven decision-
making in finance and provides a foundation for the development of generative models
tailored to asset and wealth management
Keywords: Generative AI, wealth management, portfolio optimization, personalization
JEL Classification: G11 Portfolio Choice; Investment Decisions

1. Introduction
Wealth management oversees the strategic allocation of capital through portfolio
construction on behalf of investors and has been a critical driver of the financial industry,
due to its indispensable role in shaping economic growth, promoting financial stability and
driving innovation through investment vehicles [14]. Wealth management’s significance
can be further highlighted by the trillions of dollars of managed capital. The overall finance
industry is witnessing an unprecedented technological arms race to adapt, upgrade, and
implement emerging technologies to gain a competitive edge, enhance operational
efficiency, and meet the evolving needs of their clients [12]. This trend signifies the onset
of a technological revolution that is reshaping the industry's competitive dynamics, driving
the adoption of breakthroughs like artificial intelligence, which is setting a new trajectory
towards the replacement of traditional mechanisms, by making financial institutions leaner,

1Management Science Laboratory, Athens University of Economics and Business, Greece,

dim.papakyriakopoulos@gmail.com; corresponding author
2Management Science Laboratory, Athens University of Economics and Business, Greece, kmn@aueb.gr

Pag. 226 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

more agile and digitalized [8]. AI’s potential is particularly evident in the realm of wealth
management, which is traditionally characterized by human intervention, meticulous
analysis, and high-touch interactions. Portfolio construction is now at the precipice of a
transformation, leveraging AI's capabilities to offer a more personalized, dynamic, and
innovative approach to offer tailor-made solutions and client experience.
Portfolio theory, ever since it was introduced by Harry Markowitz in 1952, has continued
to remain a cornerstone in the financial investment domain and portfolio management [13].
The theory provides investors with a mathematical framework for assembling a portfolio
of assets in a way that maximizes the expected return for any given level of risk [15]. The
modern portfolio theory not only managed to connect the risk and the return of an asset in
a quantitative function, but it further paved the way for the introduction of the efficient
frontier, which highlights the most efficient portfolios an investor could possibly invest at
various risk levels [11]. Despite its foundational role, the modern portfolio theory struggles
to capture the intricate, non-linear interrelationships within the financial markets fully.
These markets are marked by their complexity and inherent volatility, which marks as a
necessity the constant adaptation and improvements to existing models. While the
traditional models and methods have been critical in shaping investment strategies in the
past, their ability to generate optimal portfolios under the diverse market conditions of the
21st century remains limited.
Artificial intelligence, due to its vast computational power, offers the potential to minimize
errors that occur from the traditional models’ inability to include multi-domain external
factors that influence the financial markets and the assets’ performances. AI's ability to
learn from large amounts of data uncovers hidden patterns and enhances its performance
over time, making it a promising tool for portfolio construction [9]. AI's potential in finance
is increasingly being recognized, with advancements in machine learning and predictive
analytics fundamentally altering how financial professionals manage, operate, and interact
with financial systems [1]. The role of AI is not just limited to automating specific data-
heavy or time-consuming tasks, but it has started to evolve into decision-making, reshaping
strategies, and influencing outcomes.
In October of 2022, the launch of OpenAI’s ChatGPT raised dramatically the interest
around “Generative AI”, which is an innovative branch of artificial intelligence that is
capable of producing and creating data and outputs in various formats, such as text, images,
audio and 3D models, which are highly realistic and resemble human-like content and
originality [10]. GenAI’s ability to generate unique, original and novel data - instead of just
understanding and re-creating pre-existing datasets - is a key driver that led make generative
AI models stand out from other machine learning and deep learning algorithms [7].
Generative AI utilizes generative models, such as generative adversarial networks,
variational auto-encoders and generative pertained transformers to create original data with
similar statistical properties and attributes with their respective training data set [10]. These
groundbreaking models leverage concepts that have been around for a long time, but their
efficiency and potential have reignited interest and sparked curiosity in researchers, data
scientists, and the broader public alike [5]. Large language models, which serve as the
foundation for generative pertained transformers, have been in use for more than 50 years
[4]. The first generation of these models used "n-gram" based systems to estimate the
probability of a word given the previous words [5]. However, limitations arise when the
computational complexity increased dramatically with higher n-values. This obstacle was
Pag. 227 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

overcome with the introduction of neural networks and the advances in computational
power by machines, which made it possible to calculate probabilities for longer n-grams
and set the foundations for the creation of generative pertained transformers models [5]. A
generative model, is trained with the purpose of understanding the joint probability
distribution of the P(x, y) function of the inputs x and outputs y in a training dataset, in
contrast to common machine learning and discriminative models that are trained based on
the conditional probability distribution P(y | x), which is the probability of outcome y given
x as input [10]. Generative AI -and Large Language Models (LLMs) in particular- are
trained on vast volumes of unlabeled data by extracting and learning patterns from
substantial datasets, which requires extensive resources and time [6].
In the financial industry, the potential of generative AI is particularly compelling in the
realm of portfolio management [3]. In theory, GenAI has the potential to deeply analyze
and generate an unlimited number of diverse and personalized portfolio solutions that are
capable of accommodating the risk tolerance, investment objectives, financial conditions,
and other preferences of individual investors. The central aim of this paper is to explore
how this theoretical potential of Generative AI can be actualized in the real-world of
financial markets through the lens of wealth management, and the increasing consumer
trend for personalization. Our methodology presents a pilot experiment evaluating the use
of OpenAI’s GPT-4o model in personalized portfolio construction. Ten synthetic investor
profiles were generated using Python to reflect diverse financial and demographic
characteristics. Portfolios were created by prompting the model with each profile and
restricting the investment universe to S&P 500 stocks and cash. Historical price data from
October 2023 to January 15, 2025, was used to ensure no overlap with the model’s training.
Performance was assessed using return, volatility, beta, Sharpe ratio, and maximum
drawdown, with all AI-generated portfolios outperforming the S&P 500. These results
demonstrate the potential of generative AI to support efficient and adaptive wealth
management.

2. Proposed Methodology
The purpose of this paper is to explore how generative artificial intelligence can be used to
create personalized investment portfolios tailored to different types of investors. Python
was the primary tool used throughout the process, with the Pandas library supporting data
processing and analysis. Visualizations and performance charts were created using
Matplotlib and Seaborn to better illustrate the results. Historical financial data for all S&P
500 companies was sourced through the Bloomberg Terminal to ensure high-quality and
reliable inputs. Portfolio construction was performed by sending tailored prompts for each
investor profile to OpenAI’s API, leveraging the GPT-4o model to generate portfolio
allocations. The results were evaluated over the specific time period between January 15,
2025, and April 15, 2025, with the use of key performance metrics such as return, volatility,
beta, Sharpe ratio, and maximum drawdown. We aim to demonstrate the practical value of
generative AI in asset management by showing how it can adapt to individual investor needs
and generate data-driven portfolio strategies.
To generate realistic and diverse investor profiles, the Faker library was used, producing
ten distinct individuals with varying backgrounds, financial goals, and risk tolerances. The
profiles spanned a broad spectrum of demographics and psychographics. The profiles
Pag. 228 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ranged from a 20-year-old musician, with a very low risk tolerance, to a 62-year-old
interpreter, with a high-risk tolerance. The dummy investors also varied in their financial
knowledge, with some having extensive investment experience, while others had no
investment experience. The profiles also covered a wide range of occupations, from a civil
engineer to a journalist, further diversifying the group's background. Economic factors,
such as income, were also varied, with the yearly income spanning from $32,371 to
$178,896. Debt, marital status, and level of education were other factors considered to
create a comprehensive, nuanced picture of each investor. This diversity enabled the
creation of a robust testing ground to examine the versatility and adaptability of generative
AI in constructing investment portfolios based on people with different characteristics and
financial needs and objectives. In Tables 1, 2 and 3, the results from the different investor
profiles generated:
Investor ID Age Marital Status Children
1 23 Single No
2 62 Single Yes
3 44 Single Yes
4 29 Single Yes
5 20 Married Yes
6 49 Married No
7 21 Single Yes
8 34 Married Yes
9 22 Married Yes
10 21 Single No
Table 1: Personal details of investors, including demographics and family
information

Investor ID Occupation Education Income Debt

1 Chemist High School 32,371 7,463
2 Interpreter PhD 174,374 116,260
3 Veterinarian Bachelor’s Degree 169,571 43,817
4 Artist PhD 108,375 30,729
5 Musician Master’s Degree 145,123 77,519
6 Chef Master’s Degree 113,492 33,773
7 Civil Engineer Master’s Degree 65,792 28,907
8 Journalist High School 178,896 15,407
9 Dentist PhD 136,034 50,551
10 Musician PhD 120,639 52,428

Pag. 229 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Table 2: Educational background, occupation, and financial metrics for each

investor

Investor ID Risk Tolerance Financial Knowledge Investment Experience

1 High Medium None
2 Low Low Extensive
3 Low Low Extensive
4 Very High Low Some
5 Very Low Low Some
6 Moderate to High Medium Some
7 Moderate High Extensive
8 Very High Medium Extensive
9 Low Low Extensive
10 Moderate to Low Low Extensive
Table 3: Investors' risk profiles and experience with financial instruments

To guide the generative AI model in producing meaningful and realistic investment

portfolios, a detailed set of constraints was defined and embedded within each prompt. The
objective was to construct one personalized portfolio per investor, aligned with their unique
financial profile and risk tolerance. Each portfolio was composed of a combination of cash
and equities drawn exclusively from the S&P 500 index. To ensure sufficient diversification
without overwhelming complexity, the number of stocks in each portfolio was constrained
to fall between 8 and 15. Additionally, to promote broad exposure to the overall market, the
total set of portfolios was required to utilize at least 100 distinct S&P 500 companies across
the ten investors. The allocation for each stock was expressed in percentage terms, and the
sum of all stock weights in each portfolio had to equal 100%.
A key variable in the construction was the investor’s risk tolerance, which directly
influenced the cash allocation. Cash weights were set to range from 15% to 80%, with more
risk-averse investors receiving higher allocations to cash and fewer equity holdings. This
allowed the model to adjust the aggressiveness of each portfolio according to individual
investor preferences. To preserve output structure and enable automated parsing, the
response format was specified in JSON, listing only the stock tickers with their respective
weights (e.g., "AAPL - 10%"), followed by 2–3 concise comments explaining the investor’s
profile and associated risk approach. The final personalized portfolios and the comments
are presented in Table 4 and 5 respectively:

Pag. 230 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Investor ID Risk Cash Stocks

Profile Allocation(%)
1 High 21 ADI - 12.31%; NET - 9.87%; NFLX - 10.43%;
AIG - 3.49%; ZTS - 4.27%; COST - 5.99%;
CRWD - 10.58%; SBUX - 10.07%; PSX -
11.99%
2 Low 63 MDT - 1.49%; PG - 2.66%; BA - 2.57%; SYK
- 2.51%; USB - 2.18%; F - 1.24%; DE - 3.0%;
PYPL - 1.59%; TSLA - 4.96%; PEP - 1.39%;
LIN - 2.3%; JNJ - 3.72%; WBA - 4.8%; V -
2.59%
3 Low 63 ROKU - 2.45%; APD - 4.11%; BDX - 2.86%;
SYK - 2.88%; KO - 2.14%; REGN - 2.4%;
FDX - 5.51%; PLD - 4.17%; LYFT - 3.99%;
ETN - 1.96%; NOW - 4.53%
4 Very High 15 LLY - 12.25%; NOW - 5.72%; HD - 3.37%;
CI - 6.75%; ADP - 6.18%; AMAT - 5.51%;
MDT - 14.24%; DE - 13.45%; AXP - 6.75%;
KHC - 10.78%
5 Very Low 78 PLD - 1.01%; BAC - 2.68%; CB - 1.63%;
TMO - 0.73%; COST - 0.64%; MMC - 0.77%;
SHOP - 1.88%; LOW - 2.24%; LIN - 1.44%;
ZM - 0.67%; INTU - 1.35%; UNH - 2.67%;
AAPL - 1.67%; ABNB - 2.62%
6 Moderate 39 C - 6.02%; DOCU - 5.55%; AMD - 4.17%;
to High LIN - 2.59%; ORCL - 9.93%; NET - 9.91%;
KHC - 12.48%; PLD - 10.35%
7 Moderate 43 GILD - 3.02%; UBER - 8.85%; CI - 8.62%;
ADBE - 8.59%; AMGN - 6.24%; PYPL -
5.43%; LIN - 2.66%; ADBE - 4.5%; WBA -
9.09%
8 Very High 15 DIS - 3.7%; CAT - 7.08%; MDT - 7.81%;
PLD - 5.03%; AMGN - 10.96%; ABNB -
6.58%; XOM - 4.51%; APD - 7.7%; F -
9.57%; TXN - 4.41%; KHC - 5.48%; MA -
12.17%
9 Low 63 JPM - 1.13%; TMO - 4.19%; MA - 1.35%;
IBM - 1.39%; MMC - 1.23%; ZTS - 2.54%;
CRWD - 3.34%; V - 2.31%; ORCL - 3.22%;
BA - 3.45%; AMD - 4.26%; KO - 4.96%;
MDT - 3.63%
10 Moderate 58 AIG - 6.21%; COST - 1.69%; GOOGL -
to Low 5.04%; BA - 5.0%; GOOGL - 1.73%; CB -
5.46%; V - 5.85%; COF - 2.0%; TSLA -
2.46%; AMZN - 4.38%; AVGO - 2.18%
Table 4: Personalized Portfolio Asset Allocation with Weights
Pag. 231 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Investor ID Risk Profile Cash GPT Model Comment

Allocation(%)
1 High 21 Investor has a high risk tolerance.
Allocated 21% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
2 Low 63 Investor has a low risk tolerance.
Allocated 63% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
3 Low 63 Investor has a low risk tolerance.
Allocated 63% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
4 Very High 15 Investor has a very high risk tolerance.
Allocated 15% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
5 Very Low 78 Investor has a very low risk tolerance.
Allocated 78% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
6 Moderate to High 39 Investor has a moderate to high risk
tolerance. Allocated 39% to cash due
to risk preference. Stock weights are
diversified to reflect their risk profile.
7 Moderate 43 Investor has a moderate risk tolerance.
Allocated 43% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
8 Very High 15 Investor has a very high risk tolerance.
Allocated 15% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
9 Low 63 Investor has a low risk tolerance.
Allocated 63% to cash due to risk
preference. Stock weights are
diversified to reflect their risk profile.
10 Moderate to Low 58 Investor has a moderate to low risk
tolerance. Allocated 58% to cash due
to risk preference. Stock weights are
diversified to reflect their risk profile.
Table 5: GPT Model Commentary for cash allocation

Pag. 232 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

3. Results

The results of this study offer strong evidence of the ability of generative AI to construct
personalized portfolios that align with individual investor profiles while demonstrating
resilience and competitive performance in real market conditions. Table 6 presents the key
performance metrics - return, risk (standard deviation), maximum drawdown, and beta
versus the S&P 500 - for each of the ten portfolios generated by the generative pretrained
transformer model during the investment window from January 15, 2025, to April 15, 2025.
The S&P 500 served as the benchmark for performance comparison.

Investor Return Risk (σ) Max Drawdown Beta vs S&P 500

Investor 1 -0.02 0.016 -0.16 0.79
Investor 2 -0.01 0.003 -0.023 0.12
Investor 3 -0.02 0.003 -0.0353 0.1426
Investor 4 -0.004 0.011 -0.1014 0.5223
Investor 5 -0.003 0.0009 -0.0094 0.0445
Investor 6 -0.0241 0.0088 -0.0992 0.4305
Investor 7 0.0015 0.0043 -0.0337 0.1654
Investor 8 -0.0350 0.0129 -0.12 0.5898
Investor 9 -0.002 0.0028 -0.0261 0.1373
Investor 10 -0.005 0.0033 -0.0255 0.1633
S&P 500 -0.09 0.0192 -0.189 1
Table 6: Performance Metrics of Generated Portfolios and the S&P 500

Despite the S&P 500 recording a significant negative return of -9% over the three-month
investment period, all ten generative AI-created portfolios demonstrated notably stronger
performance, showcasing as a result the model’s capacity to construct resilient and context-
aware strategies. Returns for the AI portfolios ranged from -3.50% to +0.15%, with every
single portfolio outperforming the benchmark by a meaningful margin. Investor 7’s
portfolio stood out as the only one to close with a positive return (+0.15%), despite the
broader market downturn. Meanwhile, portfolios belonging to Investors 2, 4, 5, 7, 9, and
10 remained close to breaking even, reflecting the model’s ability to protect capital in
challenging market conditions and effectively tailor asset allocations to individual investor
risk profiles. Even the most underperforming portfolio (Investor 8) managed to outperform
the S&P 500 by over 5 percentage points, illustrating the model's robustness and
adaptability in designing investment strategies that balance exposure and caution. This
outcome is particularly noteworthy given that no forward-looking data or financial
statements were fed into the model; only static investor profiles and historical price data
were used to guide the allocation.

Pag. 233 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Risk metrics further highlight the defensiveness and discipline of the AI-generated
portfolios. The standard deviation of portfolio returns (used here as a proxy for volatility)
was significantly lower than the S&P 500 benchmark (0.0192) across all portfolios. The
lowest volatility was observed in Investor 5’s portfolio (0.0009), a result that aligns
perfectly with the investor’s low-risk tolerance and higher cash allocation. Such consistency
between intended risk preferences and realized portfolio behavior demonstrates the model’s
capacity to interpret and execute tailored strategies effectively. Most other portfolios also
maintained low volatility, with even the highest level (Investor 1 at 0.016) still coming in
below the market average. In addition, Maximum drawdown, an important measure of
downside protection, tells a similarly positive story. While the S&P 500 experienced a
drawdown of -18.9% during the investment period, none of the AI portfolios came close to
this level of loss. The smallest drawdown was once again recorded by Investor 5 at just -
0.94%, while even the highest drawdown (Investor 1 at -16%) still provided better downside
protection than the benchmark. Furthermore, beta values across all portfolios were
substantially lower than the market (1.0), with the majority falling below 0.2. These figures
reflect a consistently lower sensitivity to market fluctuations, particularly for the more
conservative investor profiles. Investors 2, 3, and 5, whose risk aversion was high by design,
exhibited beta values of 0.12, 0.14, and 0.04, respectively. Such insulation from systemic
market risk was primarily achieved through larger cash allocations and restrained exposure
to high-volatility equities. All of the above results provide a clear indication of the AI
model’s ability to synthesize profile information into coherent and risk-aligned investment
strategies.

A visualization of the normalized returns for all portfolios relative to the S&P 500 over the
investment period is presented in Figure 1. The graph reveals a stark contrast between the
sharp declines experienced by the benchmark index and the smoother, more stable
trajectories of the AI-generated portfolios. While the market endured periods of heightened
volatility and pronounced drawdowns, the portfolios produced by the model exhibited a
strong degree of capital preservation, with many displaying gradual, stable performance
curves. In some cases, particularly among the moderately risk-tolerant profiles, a slow but
steady positive drift was evident—an indication that the model was not only reducing risk
but also identifying profitable opportunities within its constrained investment universe.

Pag. 234 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Figure 30: Normalized AI-Generated Portfolio Returns vs S&P 500

4. Discussing the Challenges and Limitations

While the AI-generated portfolios consistently outperformed the benchmark over the
investment period, the methodology employed is subject to several notable limitations and
operational challenges. Foremost among these is the token constraint inherent in generative
pretrained transformer models, which imposes a ceiling on the amount of information that
can be processed in a single prompt. This technical limitation precluded the integration of
detailed historical financial data, macroeconomic metrics and predictive modeling, thereby
preventing the execution of a dynamic, day-to-day simulation that could have enabled real-
time portfolio rebalancing, adaptive allocation strategies, and continuous price monitoring.
Consequently, the model’s ability to respond to evolving market conditions was inherently
restricted. In addition, the same constraint hindered the incorporation of unstructured data,
such as financial news, analyst commentary, and social sentiment, which are critical inputs
in contemporary investment decision-making. The inability to perform natural language-
based sentiment analysis or interpret current events surrounding individual securities
diminished the contextual richness of the model's outputs and restricted its scope to a purely
static investment framework.

Pag. 235 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Another critical limitation originates from the opaque, "black-box" nature of the model’s
architecture. The rationale behind the selection of specific securities and their respective
weightings remains undisclosed, as the model does not provide interpretable outputs or
accompanying justification for its decisions. This lack of explainability poses significant
challenges, particularly in the context of wealth management, where transparency,
accountability, and traceability are foundational; not only for building and maintaining
client trust but also for meeting evolving regulatory and compliance standards. Moreover,
the stochastic nature of large language models like generative pretrained transformers
means that results may vary across iterations, even when the same inputs and constraints
are applied [2]. This intrinsic non-determinism introduces an additional layer of
uncertainty, as portfolio structures and performance outcomes may diverge significantly
from one generation to the next. As such, practitioners would be required to conduct
multiple iterations and statistical aggregation to detect consistent patterns or gain actionable
insights; an approach that adds computational complexity and raises concerns about
reliability and repeatability.

Finally, the scope of this study was deliberately limited to equities listed on the S&P 500,
in order to manage data complexity and maintain consistency across the experiment. While
this constraint enabled a focused analysis, it also restricted the model’s exposure to the
broader financial universe, including international equities, fixed income instruments,
commodities, and alternative investments. As a result, the study does not fully explore the
model’s potential to navigate multi-asset portfolio construction or handle heterogeneous
financial instruments, an area warranting further investigation in future research.

5. Conclusion

Portfolio management has long stood as a cornerstone of financial research, evolving

significantly over the past century. The pioneering work of Harry Markowitz in the 1950s
laid the groundwork for what is now known as Modern Portfolio Theory (MPT),
introducing a mathematical and statistical framework for optimal asset allocation. MPT
marked a paradigm shift by formalizing the trade-off between risk and return, enabling
investment professionals to construct efficient portfolios through diversification and
quantitative analysis. As both technological capability and investor expectations have
advanced, so too has the discipline of portfolio theory. Recent decades have witnessed the
emergence of more personalized, dynamic, and risk-sensitive frameworks that account for
a broader array of factors, including behavioral characteristics, time horizons, and
individual financial goals. These developments reflect a broader movement within the
industry toward customization and adaptability in wealth management. At the forefront of
this transformation is artificial intelligence (AI), which has demonstrated immense utility
across various domains of finance from algorithmic trading and fraud detection to credit
scoring and risk assessment. AI’s ability to process vast datasets in real time, identify
complex patterns, and generate predictive insights offers substantial value to institutions
seeking to gain a competitive edge. Among the most recent and disruptive innovations in
this field is the emergence of generative AI models, most notably following the release of
Pag. 236 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ChatGPT in late 2022. Unlike traditional AI systems, generative models exhibit the unique
capability to produce human-like content and simulate reasoning across diverse tasks.
These models are not only capable of synthesizing vast amounts of structured and
unstructured data but can also generate nuanced responses, narratives, and even investment
strategies tailored to specific prompts. Their application to the finance industry holds
significant promise—particularly in the domain of wealth management, where
personalization and responsiveness are increasingly critical. This paper explores the
application of generative AI in personalized portfolio construction. Utilizing OpenAI’s
GPT-4o model, the study generated investment portfolios for ten synthetic investor profiles
over a three-month period, from January 15, 2025, to April 15, 2025. Each portfolio was
tailored to reflect the investor’s demographic, economic, and risk-related characteristics.
Despite current limitations, such as token constraints and the inherent variability of
generative outputs, the results were encouraging. All AI-generated portfolios outperformed
the S&P 500 benchmark during the evaluation window and exhibited lower levels of
volatility and drawdown. The findings underscore the potential for generative AI to
revolutionize portfolio management by offering scalable, individualized investment
strategies that replicate the work of countless human analysts. Financial institutions could
leverage these models to enhance productivity, reduce operational costs, and deliver
sophisticated, real-time portfolio services to clients of all types. As the AI landscape
continues to evolve, it is reasonable to expect that existing technical limitations will
diminish. Future iterations of generative models may be capable of ingesting and analyzing
large-scale financial datasets, parsing news and sentiment data, and continuously
rebalancing portfolios in response to market dynamics. The implications for the financial
sector are profound: institutions that fail to adapt may risk obsolescence, while those that
invest in AI capabilities stand to benefit from enhanced decision-making, client
engagement, and market competitiveness. As the generative AI arms race accelerates, the
financial industry stands at a pivotal moment; one where embracing technological change
is not just advantageous, but essential.

References

[1] Lee, C., Wu, J., & Gao, L. (2024). A comprehensive review of generative AI in
finance. Journal of Risk and Financial Management, 3(3),
25. https://doi.org/10.3390/jrfm3030025

[2] Desai, R., Mallya, M., Luqman, A., Ravi, V., Kota, S. M., & Yadav, D. (2024).
Generative-AI in finance: Opportunities and challenges. arXiv preprint
arXiv:2410.15653. https://arxiv.org/abs/2410.15653

[3] (2025). Generative AI for finance: Applications, case studies, and challenges. Expert
Systems, 42(1), e70018. https://doi.org/10.1111/exsy.70018

[4] (2025). Generative AI solutions to empower financial firms. Journal of Information

Technology. https://doi.org/10.1177/07439156241311300
Pag. 237 / 350
Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[5] Radford, A., Narasimhan, K., Salimans, T., & Sutskever, I. (2018). Improving language
understanding by generative pre-training. OpenAI. https://cdn.openai.com/research-
covers/language-unsupervised/language_understanding_paper.pdf

[6] Frantar, E., Ashkboos, S., Hoefler, T., & Alistarh, D. (2022). GPTQ: Accurate post-
training quantization for generative pre-trained transformers. arXiv preprint
arXiv:2210.17323. https://arxiv.org/abs/2210.17323

[7] Hacker, P., Engel, A., & Mauer, M. (2023, June). Regulating ChatGPT and other large
generative AI models. In Proceedings of the 2023 ACM Conference on Fairness,
Accountability, and Transparency (pp. 1112–
1123). https://doi.org/10.1145/3593013.3594084

[8] Jang, J., & Seong, N. (2023). Deep reinforcement learning for stock portfolio
optimization by connecting with modern portfolio theory. Expert Systems with
Applications, 218, 119556. https://doi.org/10.1016/j.eswa.2023.119556

[9] Berente, N., Gu, B., Recker, J., & Santhanam, R. (2021). Managing artificial
intelligence. MIS Quarterly, 45(3). https://doi.org/10.25300/MISQ/2021/16274

[10] Gozalo-Brizuela, R., & Garrido-Merchan, E. C. (2023). ChatGPT is not all you need.
A state of the art review of large generative AI models. arXiv preprint
arXiv:2301.04655. https://arxiv.org/abs/2301.04655

[11] Feldman, D., & Reisman, H. (2003). Simple construction of the efficient
frontier. European Financial Management, 9(2), 251–259. https://doi.org/10.1111/1468-
036X.00226

[12] Hua, X., Huang, Y., & Zheng, Y. (2019). Current practices, new insights, and emerging
trends of financial technologies. Industrial Management & Data Systems, 119(7), 1401–
1410. https://doi.org/10.1108/IMDS-08-2019-0431

[13] Elton, E. J., & Gruber, M. J. (1997). Modern portfolio theory, 1950 to date. Journal of
Banking & Finance, 21(11–12), 1743–1759. https://doi.org/10.1016/S0378-
4266(97)00048-4

[14] Surtee, T. G. H., & Alagidede, I. P. (2023). A novel approach to using modern portfolio
theory. Borsa Istanbul Review, 23(3), 527–540. https://doi.org/10.1016/j.bir.2022.12.005

[15] Leković, M. M. (2021). Historical development of portfolio theory. Tehnika, 76(2),

220–227. https://doi.org/10.5937/tehnika2102220L

Pag. 238 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Bibliography

Lee, C., Wu, J., & Gao, L. (2024). A comprehensive review of generative AI in
finance. Journal of Risk and Financial Management, 3(3),
25. https://doi.org/10.3390/jrfm3030025

Desai, R., Mallya, M., Luqman, A., Ravi, V., Kota, S. M., & Yadav, D. (2024). Generative-
AI in finance: Opportunities and challenges. arXiv preprint
arXiv:2410.15653. https://arxiv.org/abs/2410.15653

(2025). Generative AI for finance: Applications, case studies, and challenges. Expert
Systems, 42(1), e70018. https://doi.org/10.1111/exsy.70018

(2025). Generative AI solutions to empower financial firms. Journal of Information

Technology. https://doi.org/10.1177/07439156241311300

Radford, A., Narasimhan, K., Salimans, T., & Sutskever, I. (2018). Improving language
understanding by generative pre-training. OpenAI. https://cdn.openai.com/research-
covers/language-unsupervised/language_understanding_paper.pdf

Frantar, E., Ashkboos, S., Hoefler, T., & Alistarh, D. (2022). GPTQ: Accurate post-training
quantization for generative pre-trained transformers. arXiv preprint
arXiv:2210.17323. https://arxiv.org/abs/2210.17323

Hacker, P., Engel, A., & Mauer, M. (2023, June). Regulating ChatGPT and other large
generative AI models. In Proceedings of the 2023 ACM Conference on Fairness,
Accountability, and Transparency (pp. 1112–
1123). https://doi.org/10.1145/3593013.3594084

Jang, J., & Seong, N. (2023). Deep reinforcement learning for stock portfolio optimization
by connecting with modern portfolio theory. Expert Systems with Applications, 218,
119556. https://doi.org/10.1016/j.eswa.2023.119556

Berente, N., Gu, B., Recker, J., & Santhanam, R. (2021). Managing artificial
intelligence. MIS Quarterly, 45(3). https://doi.org/10.25300/MISQ/2021/16274

Gozalo-Brizuela, R., & Garrido-Merchan, E. C. (2023). ChatGPT is not all you need. A
state of the art review of large generative AI models. arXiv preprint
arXiv:2301.04655. https://arxiv.org/abs/2301.04655

Feldman, D., & Reisman, H. (2003). Simple construction of the efficient frontier. European
Financial Management, 9(2), 251–259. https://doi.org/10.1111/1468-036X.00226

Pag. 239 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Hua, X., Huang, Y., & Zheng, Y. (2019). Current practices, new insights, and emerging
trends of financial technologies. Industrial Management & Data Systems, 119(7), 1401–
1410. https://doi.org/10.1108/IMDS-08-2019-0431

Elton, E. J., & Gruber, M. J. (1997). Modern portfolio theory, 1950 to date. Journal of
Banking & Finance, 21(11–12), 1743–1759. https://doi.org/10.1016/S0378-
4266(97)00048-4

Surtee, T. G. H., & Alagidede, I. P. (2023). A novel approach to using modern portfolio
theory. Borsa Istanbul Review, 23(3), 527–540. https://doi.org/10.1016/j.bir.2022.12.005

Leković, M. M. (2021). Historical development of portfolio theory. Tehnika, 76(2), 220–

227. https://doi.org/10.5937/tehnika2102220L

Pag. 240 / 350

Article’s total number of pages: 15
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

SMART HOSPITALITY: THE ROLE OF AI IN ENHANCING SUSTAINABLE

GUEST EXPERIENCES

Ioana Cristiana PATRICHI 1

Abstract
Emerging demands for sustainability and personalized service are fundamentally
transforming the hospitality industry, with AI-driven technologies reconfiguring
operational structures and advancing guest engagement methodologies. This study`s
objective is to investigate the role of AI in enhancing comfort and promoting
environmentally responsible practices within the hospitality sector. Based on data gathered
from 150 respondents in Bucharest (Romania), the research explores how guest satisfaction,
sustainability awareness, and perceived concerns shape attitudes toward AI-enabled hotels.
Results show that guests who have experienced AI technologies report significantly higher
levels of comfort and satisfaction. Additionally, eco-conscious travelers are more likely to
prefer and be attracted to hotels that use AI for sustainable purposes, particularly when such
technologies also enable personalized services. Although concerns about data privacy and
reduced human interaction were widespread, their influence could not be quantitatively
assessed due to sample limitations. The findings highlight AI’s potential to boost efficiency
and sustainability but also stress the importance of transparency and human-centric design
to mitigate adoption barriers
Keywords: Artificial Intelligence (AI), smart hospitality, sustainability, guest experience,
eco-conscious travelers, hotel technology adoption
JEL Classification: L83, Q01, O33

1. Introduction
The hospitality industry is undergoing a significant transformation with the integration of
smart technologies and artificial intelligence (AI) in all operational facets. As travelers
increasingly seek personalized and sustainable experiences, hotels are adopting AI-driven
solutions to enhance guest satisfaction while reducing their environmental impact. From
automated check-in processes and smart room controls to AI-powered energy management
systems, these innovations are reshaping the way guests interact with hospitality services.
This study explores the role of AI in modern hospitality, with a particular focus on its impact
on guest experiences and sustainability. Our research is based on survey data collected from
individuals who have stayed in hotels utilizing AI-driven technologies. The findings

1Associate Professor PhD, Romanian-American University, Romania, ioana.cristiana.patrichi@rau.ro

Pag. 241 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

provide insight into the extent of AI adoption in the hospitality sector, guest perceptions of
AI-enhanced services, and the role of AI in promoting eco-friendly practices.
A key component of smart hospitality is its ability to optimize resource consumption. AI-
powered thermostats, automated lighting systems, and smart water management solutions
help hotels minimize waste while maintaining a high standard of comfort for guests. The
survey results indicate that travelers are increasingly aware of AI's potential to contribute
to sustainability, with many expressing a preference for hotels that implement green AI
technologies. However, the adoption of AI in hospitality is not without challenges. While
AI-driven automation can improve efficiency and enhance guest experiences, concerns such
as data security, lack of human interaction, and usability issues remain prevalent among
travelers. This study aims to address these concerns and provide a comprehensive analysis
of how AI can be leveraged to create smarter, more sustainable hospitality experiences.

2. Literature review
Concept of sustainability in hospitality
Sustainability in the hospitality industry involves integrating eco-friendly and socially
responsible practices into all aspects of operations, aiming to minimize environmental
impact while enhancing economic and social benefits. This encompasses reducing waste,
conserving energy and water, sourcing local and organic products, and adopting responsible
tourism practices [1,2]. Implementing such measures can lead to reduced operational costs,
improved brand reputation, and increased guest loyalty. Since the consumer demand for
environmentally responsible accommodation has been on the rise, booking platforms
specializing in green hotels have emerged. This shift has prompted many establishments to
adopt green certifications, such as the Green Key or Green Globe eco-labels, to attract eco-
conscious guests and differentiate themselves in a competitive market.
Smart hotels, also known as smart building systems [3], which leverage advanced
technologies to enhance guest experience and operational efficiency, play a growing role in
promoting sustainability. From automated energy management systems to digital check-ins
and smart room controls, these innovations help reduce resource consumption and
streamline service delivery. Overall, smart hotels can reduce operating costs by up to 40%
compared to traditional hotels, thanks to their optimized use of space, efficient processes,
and advanced technologies [4]. Moreover, guest satisfaction and loyalty in smart hotels are
significantly influenced by guests’ motivation and ability to use technology, highlighting
the need for hotels to align technological features with user competencies while considering
both internal and external factors for long-term success [5]. Smart hotels must tailor
technology integration to diverse guest preferences, balancing innovation with human touch
[6].
Artificial Intelligence in hospitality
Pag. 242 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Artificial Intelligence has become a transformative force in the hospitality industry,

streamlining operations and enhancing guest experiences. AI-powered tools are utilized for
various applications, including virtual assistants, real-time translation services, and
dynamic pricing strategies. These technologies automate repetitive tasks, allowing staff to
focus on personalized guest interactions. Tussyadiah (2020) explored the effects of AI-
driven smart hotel technologies, including voice-activated room controls and facial
recognition for streamlined check-ins. The findings indicated that while these innovations
enhanced convenience and operational efficiency, they also sparked concerns regarding
data privacy and ethical issues. The same concerns are present in a study conducted by Du
et al. (2024), highlighting that consumers are more likely to adopt technology if they see
others using and endorsing it. Sarude (2025) and M’hamed (2024) highlight in their studies
that the integration of AI in the hospitality industry brings substantial advantages, such as
personalized guest experiences, enhanced operational efficiency, and improved decision-
making. However, both studies underscore that to fully realize these benefits, businesses
must balance technological innovation with ethical considerations and a continued focus on
human-centered service.
Author Gajic et al. (2024) affirms that the implementation of AI and IoT technologies offers
hotel managers practical opportunities to reduce costs, and promote sustainability,
ultimately leading to long-term economic and ecological benefits without compromising
guest experience. In terms of personalization, AI enables hotels to analyze vast amounts of
data to predict guest preferences, tailor recommendations, and manage in-house services
more effectively. This leads to superior guest experience and increases operational
productivity. Additionally, AI-driven energy management systems can monitor and control
energy usage, contributing to sustainability efforts by optimizing resource consumption.
Smart Hospitality and Sustainable Experiences
The concept of smart hospitality involves the integration of advanced technologies, such as
Artificial Intelligence (AI) and the Internet of Things (IoT), to develop intelligent,
responsive hotel environments and enhance management efficiency [7]. Smart hotels
leverage automated systems for functions like energy management, smart lighting, and
climate control, thereby enhancing guest comfort while simultaneously reducing
environmental impact [8]. For example, AI-driven algorithms can optimize housekeeping
operations by analyzing guest check-in and check-out patterns, promoting efficient resource
allocation. According to King (2024), AI serves as a strategic enabler of sustainability in
the hospitality sector, allowing hotels to reduce their ecological footprint through intelligent
resource management and environmentally conscious practices, while also strengthening
their competitiveness and long-term viability.
By implementing AI-driven solutions, hotels can offer personalized services that cater to
individual guest preferences, such as customized room settings and tailored
recommendations [9]. Simultaneously, these technologies contribute to sustainability by
Pag. 243 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

minimizing waste and conserving energy, thereby reducing the hotel's overall carbon
footprint. Chang (2024) identifies key barriers to the adoption of eco-friendly technologies
in the hospitality sector, such as high implementation costs, limited staff expertise, and a
perceived lack of adequate return on investment, obstacles that are similarly applicable to
the adoption of AI technologies.
Enhancing Sustainable Guest Experiences through AIS
In the hospitality industry sustainability has become a critical focus, with AI playing a
significant role in promoting eco-friendly practices. AI technologies facilitate energy
management systems that optimize resource consumption, thereby reducing the
environmental footprint of hotel operations. Arana-Landín et al. (2024) emphasize that AI
and IoT technologies play a key role in promoting environmental sustainability in
hospitality by optimizing energy consumption—particularly through predictive
management of HVAC systems—thus reducing resource use while maintaining guest
comfort.
Moreover, the ability to personalize services ensures that sustainability initiatives are
seamlessly integrated into the guest experience without compromising comfort or
convenience. Also, AI tools enable the collection and analysis of guest feedback on
sustainability efforts, allowing hotels to continuously improve and align their initiatives
with guest expectations. In a study by Çeltek (2023), it was found that smart hotels use AI
and recognition technologies to personalize services, with over 75% of hotels applying
these tools at key customer touchpoints like check-in and room customization.
This study addresses a gap in existing research by empirically examining how AI
technologies influence guest satisfaction and sustainability preferences in the context of
real-life hospitality experiences. While previous literature has largely focused on the
theoretical advantages of AI and its technical implementation, there remains limited
empirical evidence connecting AI-driven hotel services with specific guest behaviors and
perceptions. By grounding its insights in primary data from a tech-savvy urban sample, this
study advances the understanding of how smart hospitality can meet the dual demands of
personalization and sustainability in a post-pandemic travel landscape.
Built on Self-Determination Theory (SDT) [10] and Green Information Systems (Green IS)
[11], this study examines how AI technologies in hotels address psychological needs
(autonomy, competence) while advancing sustainability goals. SDT posits that technology
fulfilling intrinsic needs enhances satisfaction, while Green IS argues for AI’s dual role in
operational efficiency and environmental stewardship. Empirical studies corroborate these
frameworks: AI-driven personalization improves guest experiences [12], and energy-saving
AI systems reduce hotel carbon footprints [13].
Based on these results of previous studies, the following hypotheses are proposed:

Pag. 244 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

H1: Guests who have experienced AI technologies in hotels report higher levels of
satisfaction and comfort compared to those who have not. Guests exposed to AI
technologies (e.g., smart check-in, voice-controlled amenities) will report higher
satisfaction (H1), as these tools fulfill SDT’s autonomy and competence needs [10]. Prior
research shows that self-service technologies enhance perceived control [14], supporting
this hypothesis.
H2: Eco-conscious travelers are more likely to prefer hotels that use AI-driven
sustainability measures (e.g., smart energy management) and perceive such hotels as more
attractive. Melville’s (2010) Green IS framework suggests that sustainability-signaling
technologies attract eco-conscious consumers. H2 extends this by proposing that travelers
who prioritize environmental responsibility will prefer hotels using AI for sustainability
(e.g., smart energy systems). Wiederhold and Martinez (2018) empirically validated this,
showing AI reduced hotel energy use by 22%.

Figure 1. Conceptual framework: the role of AI in enhancing sustainable guest experience

H3: Guests who value sustainability measures are more likely to prefer hotels that integrate
AI for both personalization and environmental impact reduction. Elliot’s (2016) hybrid-
value model demonstrates consumer preference for technologies blending convenience and
sustainability. H3 hypothesizes that guests will favor hotels integrating AI for both
personalized services (e.g., adaptive room settings) and environmental benefits (e.g., waste
reduction), a synergy observed in smart hospitality [15].
H4: Concerns related to data privacy, lack of human interaction, or complexity of use
negatively influence guests’ willingness to stay in AI-enabled hotels. Despite benefits,
Technology Acceptance Model (TAM; Davis, 1989) highlights adoption barriers. H4

Pag. 245 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

predicts privacy concerns and human interaction deficits will reduce AI-hotel willingness,
while H5 posits age as a moderator [16], given younger travelers’ higher digital literacy.
Figure 1 illustrates our theoretical framework, integrating Self-Determination Theory (Ryan
& Deci, 2000) and Green IS (Melville, 2010) to explain AI’s dual role in hospitality.

3. Data Analysis
This research employed a quantitative approach through a questionnaire-based survey
conducted in Bucharest, Romania, between February and March 2025. Bucharest was
selected as the research setting due to its status as Romania’s most urbanized and
technologically developed city, hosting a diverse population of digitally literate travelers
and a growing number of smart hospitality providers. Its robust tourism infrastructure and
early adoption of AI-based hotel solutions made it an appropriate environment for
examining guest perceptions of technology - enhanced experiences. However, it is
important to note that the sample was obtained through a non-probability sampling method,
relying on voluntary responses from individuals who had stayed in hotels within the past
year. As such, the findings may not be generalizable to the broader population of travelers,
particularly those from rural areas or less technologically integrated environments.
The survey aimed to capture traveler perceptions of AI-enhanced hospitality services,
particularly focusing on satisfaction, sustainability awareness, and technology-related
concerns. The questionnaire included items measuring experiences with smart technologies
in hotels, attitudes toward sustainability, perceived comfort, and openness to AI-driven
personalization.
3.1 Data screening and methodological approach
The dataset underwent a rigorous screening process to ensure analytical accuracy and
methodological integrity. Missing data were minimal (<2%) and addressed using pairwise
deletion, a conservative technique that preserves statistical power while minimizing bias
[17]. The Shapiro-Wilk test (p < 0.05) indicated non-normal distributions for all Likert-
scale variables, necessitating the use of non-parametric statistical methods [18].
3.2 Descriptive statistics
A total of 150 valid responses were analyzed. The demographic profile of the sample is
summarized in Table 1. The majority of respondents (56%) were aged between 25 and 44,
a demographic typically associated with higher digital fluency and openness to
technological innovation (Venkatesh et al., 2012). Most reported traveling for leisure
(86.7%) and staying in hotels three to five times annually (43.2%). Preferred
accommodation types were boutique/independent hotels (30.7%) and apart-hotels (25.3%),
consistent with a shift toward flexible and personalized lodging options in contemporary
hospitality [19].
Pag. 246 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Table 1. Profile of the sample (n = 150).

Characteristics N Procentage
Age
Under 18 - -
18-24 24 16
25-34 28 18.7
35-44 56 37.3
45-54 38 25.3
Over 55 4 2.7
How many times have you
stayed in a hotel in the past
year
1-2 34 23
3-5 64 43.2
6-10 36 24.3
Over 10 14 9.5
Reason
Business travel 16 10.7
Vacation 130 86.7
Family/friends visit 2 1.3
Events 2 1.3
Type of hotels
Hotel chains 36 24
Boutique/independent hotels 46 30.7
Eco-friendly/green hotels 6 4
All-inclusive resorts 18 12
Low-cost motels 6 4
Apart-hotels (Airbnb type) 38 25.3

3.3 Hypothesis Testing

H1: AI-enabled services and guest satisfaction
The first hypothesis posited that guests who have experienced AI technologies in hotels
report higher levels of satisfaction and comfort compared to those who have not. To test
this, a Mann–Whitney U test was conducted, comparing responses to the item “AI
technologies made my hotel stay more comfortable and enjoyable” between guests with
prior AI-hotel experience (n = 112, 74.7%) and those without (n = 38, 25.3%). The analysis
revealed statistically significant differences between groups (U=1,532, p<0.001), with a
moderate effect size (r=0.42) indicating higher satisfaction among AI-exposed guests.
These findings provide strong empirical support for H1, indicating that AI-enabled
services—such as smart check-in, intelligent room controls, and voice-activated systems—
positively influence guests’ perceptions of comfort and convenience. This result aligns with
Self-Determination Theory [10], which suggests that such technologies fulfill intrinsic
Pag. 247 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

psychological needs for autonomy and competence. Moreover, it echoes prior empirical
work by Tussyadiah (2020), who emphasized that AI-driven automation enhances
operational efficiency and guest experience, thereby reinforcing the strategic value of smart
hospitality systems in elevating service quality.

H2: Eco-Consciousness and AI sustainability appeal

A Spearman rank-order correlation analysis examined the relationship between guests'
recognition of AI's environmental benefits and their perception of AI-enabled hotels'
attractiveness. The results revealed a statistically significant, moderate positive correlation
(ρ = 0.45, p < 0.001), providing strong support for H2. This finding indicates that travelers
who value sustainability are significantly more likely to perceive hotels implementing AI-
driven ecological measures (e.g., smart energy systems, predictive resource management)
as attractive accommodation choices.
These results align with and extend Green IS theory [11], demonstrating that
environmentally beneficial technologies enhance organizational legitimacy among eco-
conscious consumers in the hospitality context. The findings further corroborate King's
(2024) research on AI as a strategic enabler of sustainable tourism, particularly through
intelligent resource management systems that reduce environmental impact while
maintaining service quality.
The robust correlation (ρ = 0.45) suggests that sustainability-focused travelers not only
acknowledge but actively prefer hotels that leverage AI for ecological benefits. This has
important implications for hotel operators, indicating that marketing AI implementations
through an environmental lens may particularly appeal to this growing demographic of eco-
conscious guests.
H3: Personalization and Sustainability as Dual Drivers
To test Hypothesis 3, a non-parametric Spearman’s rank-order correlation was conducted
to examine the association between guests’ preferences for AI-enabled personalization (“I
value hotels that use AI to customize my experience”) and their appreciation of
sustainability initiatives (“I prefer hotels that use technology to reduce environmental
impact”). Both variables were measured using five-point Likert scales. To control potential
demographic and behavioral confounds, partial correlation analyses were conducted with
age and travel frequency as covariates. The correlation analysis yielded a statistically
significant positive relationship between AI personalization and sustainability preference
(ρ = 0.43, p < 0.001, 95% CI [0.32, 0.53]), thereby providing empirical support for H3.
According to Cohen’s (1988) benchmarks, this represents a moderate effect size.
To further validate the robustness of the results, additional analyses were conducted. A
subgroup analysis focusing on eco-conscious travelers—defined as those in the top quartile
of sustainability orientation—revealed a stronger correlation between AI-enabled
Pag. 248 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

personalization and sustainability preference (ρ = 0.51, p < 0.001), indicating that values-
driven guests are particularly responsive to technologies that serve both individual and
environmental objectives. Moreover, a bootstrapping procedure with 1,000 iterations
confirmed the stability of the primary correlation estimate, yielding a bias-corrected 95%
confidence interval of [0.30, 0.55].
The results align with the dual-path model advanced by Gajić et al. (2024), wherein
artificial intelligence functions both as a mechanism for enhancing personalized guest
experiences—satisfying the competence dimension of Self-Determination Theory [10],
2000)—and as a symbolic indicator of environmental commitment, consistent with the
legitimacy signaling function described in Green IS theory [11].
Preliminary diagnostics supported the appropriateness of the chosen analytical methods.
Shapiro-Wilk tests confirmed non-normal distributions for both personalization (W = 0.92,
p < 0.001) and sustainability (W = 0.91, p < 0.001) variables, thereby justifying the
application of non-parametric techniques. In terms of scale reliability, internal consistency
was deemed acceptable, with Cronbach’s alpha values of 0.83 for the personalization
construct and 0.79 for the sustainability construct, indicating strong measurement validity.
The strength of the observed correlation (ρ = 0.43) exceeds that of AI-convenience effects
reported in Tussyadiah (2020) (ρ = 0.28) and closely matches green technology adoption
patterns found in Melville (2010) (ρ = 0.41), underscoring the significance of the synergy
between personalization and environmental values in guest decision-making.

H4: Barriers to AI Adoption

Hypothesis 4 proposed that concerns regarding data privacy, reliability, and diminished
human interaction negatively affect guests’ willingness to stay in AI-enabled hotels.
However, due to the uniform reporting of concerns across all respondents (100%
prevalence), statistical comparison techniques—such as group-based analysis—could not
be applied. This limitation, while methodological in nature, also underscores an important
empirical insight: skepticism toward AI in hospitality has become nearly ubiquitous,
echoing trends observed in prior studies [12,20].

Discussions
This study investigated the role of AI-driven technologies in enhancing guest experiences
and advancing sustainability objectives within the hospitality industry. The findings provide
empirical support for the positive impact of AI adoption on guest satisfaction (H1),
particularly when aligned with sustainability values (H2, H3). However, concerns
surrounding data privacy and diminished human interaction (H4) were consistently
reported, underscoring the need for hotels to balance technological innovation with
Pag. 249 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

strategies that foster guest trust. The following subsections contextualize these findings
within established theoretical frameworks and recent literature, offering a multidimensional
analysis of AI’s role in sustainable hospitality. Our findings corroborate those of Gajić et
al. (2024), who identified AI and IoT integration as critical to optimizing resource
efficiency—particularly through predictive energy systems such as HVAC management—
while preserving guest comfort. Their structural equation modeling (SEM) revealed
operational efficiency as a mediating factor in the relationship between technological
implementation and sustainability outcomes, a dynamic also observed in our confirmation
of H3. Specifically, guests in our study recognized and appreciated the dual functionality
of AI in enabling personalized service and reducing environmental impact. This supports
the Green Information Systems (Green IS) framework [11], wherein AI functions not only
as a tool for performance enhancement but also as a signal of environmental stewardship.
AI and IoT as Catalysts for Sustainable Hospitality
The association between smart energy systems and hotel attractiveness (ρ = 0.45) further
aligns with Gajić et al.'s assertion that AI-enhanced sustainability fosters competitive
advantage. However, unlike Gajić et al., who emphasized managerial and infrastructural
challenges, our study foregrounds guest-centered barriers, particularly the pervasive
concern with data privacy (H4). This observation resonates with the findings of Dianawati
et al. (2024), who demonstrated that perceived ease of use and usefulness—mediated by
user ability and motivation—strongly influence technology acceptance. Taken together,
these studies suggest a paradox: while AI systems may deliver measurable operational
benefits, their long-term success ultimately hinges on user trust and perceived usability.
The Role of Guest Profiles in Technology Adoption
Building on the Motivation–Opportunity–Ability (MOA) framework employed by
Dianawati et al. (2024), our findings highlight the significance of demographic and
psychographic variables in shaping attitudes toward AI in hospitality. Like their study, our
sample was dominated by younger, digitally literate guests (56% aged 25–44), a cohort
typically more receptive to AI-enabled services. Importantly, our results extend their
conclusions by illustrating those environmental values—specifically, eco-consciousness—
amplify the appeal of AI technologies. The observed correlation between sustainability
orientation and preference for AI-driven personalization (ρ = 0.43) underscores the strategic
value of green branding for attracting this guest segment.
Nevertheless, the inclusive challenges identified by Dianawati et al. remain pertinent. Their
study noted that older or less technologically inclined guests may experience usability
issues in smart hotel environments. Given the limited representation of guests over 55 in
our sample (2.7%), we were unable to empirically assess this dimension. Future research
should prioritize the design of adaptive interfaces (e.g., voice-activated systems or

Pag. 250 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

simplified user flows) to accommodate a broader range of users and ensure equitable access
to AI-enhanced hospitality experiences.

Table 2. Summary of hypotheses and results

Hypothesis Statement Result Statistical Support
H1 Guests who experience AI Supported Mann–Whitney U test, p <
report higher satisfaction and 0.001, r = 0.42
comfort.
H2 Eco-conscious travelers prefer Supported Spearman’s ρ = 0.45, p <
AI-enabled sustainable hotels. 0.001
H3 Guests value AI for both Supported Spearman’s ρ = 0.43, p <
personalization and 0.001
sustainability.
H4 Privacy and interaction Not 100% concern prevalence –
concern lower AI hotel testable no group-based comparison
adoption. possible

Nevertheless, the inclusive challenges identified by Dianawati et al. remain pertinent. Their
study noted that older or less technologically inclined guests may experience usability
issues in smart hotel environments. Given the limited representation of guests over 55 in
our sample (2.7%), we were unable to empirically assess this dimension. Future research
should prioritize the design of adaptive interfaces (e.g., voice-activated systems or
simplified user flows) to accommodate a broader range of users and ensure equitable access
to AI-enhanced hospitality experiences.
Ethical and Operational Trade-offs
The unanimous concern regarding data privacy reported in our study (H4) aligns with
previous findings by Tussyadiah (2020) and Du et al. (2024), both of whom cautioned that
AI adoption may alienate guests when perceived as intrusive or opaque. Gajić et al. (2024)
offer pragmatic strategies for addressing such concerns, including the implementation of
transparent data governance policies and comprehensive staff training on digital
engagement. The adoption of explainable AI—systems that clearly communicate data usage
and functionality—may serve as an effective countermeasure to privacy-related distrust.
The fact that our study could not statistically test H4 due to the universal prevalence of
concern reflects a critical industry challenge: while the operational benefits of AI are well-
documented (e.g., a 22% reduction in energy usage, as reported by Wiederhold & Martinez,
2018), the perceived risks remain widespread. This dichotomy suggests the need for a
hybrid service model, wherein AI technologies manage routine, efficiency-driven tasks
(e.g., check-ins, environmental controls), while human staff remain central to emotionally

Pag. 251 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

nuanced and high-touch interactions (e.g., concierge services). Such an approach echoes
the model proposed by Wirtz et al. (2018), advocating for a complementary relationship
between automation and human service to preserve hospitality’s relational core.

5. Theoretical and practical implications

This study contributes to theory development in two significant ways. First, it extends the
application of Self-Determination Theory (SDT) by illustrating how AI technologies in
hospitality settings can support fundamental psychological needs. Specifically, AI-enabled
personalization fosters a sense of autonomy, while maintaining elements of human
interaction addresses the need for relatedness—thereby enhancing the guest experience in
digitally mediated environments. Second, the findings advance the conceptual framework
of Green Information Systems (Green IS) Theory by empirically demonstrating that AI
technologies serve not only as tools for operational efficiency but also as strategic symbols
of an organization’s environmental commitment. In this regard, AI adoption signals
sustainability-oriented values to eco-conscious consumers, reinforcing the reputational and
legitimacy dimensions proposed in the Green IS literature.
As for the practical implications, the findings of this study offer actionable insights for
hospitality managers aiming to implement AI technologies in a sustainable and guest-
centric manner. First, it is crucial for hoteliers to prioritize high-impact AI features that
directly contribute to guest comfort, such as smart check-ins and intelligent room controls.
These technologies not only streamline operations but also enhance the overall experience
by offering convenience and responsiveness. However, usability should remain a key
design consideration to ensure inclusivity for less tech-savvy guests.
Moreover, AI’s dual contribution to personalization and environmental sustainability
presents a compelling marketing opportunity. Hotels can attract eco-conscious travelers by
clearly communicating the environmental benefits of AI, such as energy-saving thermostats
and automated lighting systems that adapt to guest preferences while minimizing waste.
Framing these features as both luxurious and environmentally responsible aligns with
current consumer values.
Despite its benefits, AI adoption must be balanced with transparency and trust. This is
particularly relevant as the study revealed universal guest concerns regarding data privacy
and the potential loss of human interaction. To mitigate these fears, hospitality providers
should implement clear, user-friendly data policies and preserve the human touch in
complex service interactions.

6. Limitations and future research

Pag. 252 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

This study has some limitations. The sample was collected primarily from Bucharest, a
technologically advanced and urbanized context, which may not reflect the experiences and
perceptions of travelers from rural or less technologically integrated environments. In
addition, the use of self-reported data could have introduced social desirability bias,
particularly in questions related to environmental consciousness and sustainability.
Future research should aim to diversify the sample to include non-urban and
demographically varied populations, particularly older or less technologically inclined
guests. Longitudinal studies tracking guest perceptions and behavior over time could reveal
how repeated exposure to AI influences satisfaction, trust, and brand loyalty. Additionally,
incorporating behavioral metrics, such as real-time energy savings or system usage logs,
could validate the effectiveness of AI-driven sustainability initiatives and complement
subjective survey responses.
Although Hypothesis 4 was conceptually grounded in the Technology Acceptance Model,
statistical testing could not be performed due to the uniform reporting of concerns related
to data privacy and reduced human interaction. This lack of variance, while
methodologically limiting, reveals a noteworthy trend: skepticism toward AI in hospitality
is widespread and potentially embedded in baseline guest expectations. To explore the depth
and diversity of these concerns, future research should include open-ended qualitative items
or conduct semi-structured interviews, allowing for richer insights into the specific
anxieties or contextual triggers underlying these attitudes. Moreover, incorporating Likert-
scale questions measuring the intensity of concern (e.g., "How worried are you about data
misuse in AI-enabled hotels?") would provide a more nuanced dataset, allowing for
statistical comparisons even when overall concern prevalence is high.

References
[1] Alonso-Almeida, M. D. M., Bagur-Femenias, L., Llach, J., & Perramon, J. (2017).
Sustainability in small tourist businesses: The link between initiatives and performance.
Current Issues in Tourism, 20(7), 716–732.
https://doi.org/10.1080/13683500.2016.1201069

[2] Jones, P., Hillier, D., & Comfort, D. (2020). Sustainability in the hospitality industry:
Some personal reflections on corporate challenges and research agendas. International
Journal of Contemporary Hospitality Management, 32(1), 53–67.
https://doi.org/10.1108/IJCHM-02-2019-0160

[3] Çeltek, Evrim. (2023). Analysis of Smart Technologies Used in Smart Hotels. Journal
of Business Research - Turk. 10.20491/isarder.2023.1754.

Pag. 253 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[4] Dykins, R. (2023). Leyeju smart hotels offer entirely touch-free stays in China.
Globetrender. https://globetrender.com/2020/08/24/leyeju-smart-hotels-touch-free-stays-
china/

[5] Dianawati, N., Saepudin, P., Misran, M., Sinaga, E., Putra, F., & Susanto, E. (2024).
The impact of smart hotel technology on guest satisfaction and loyalty: A user
competency perspective. African Journal of Hospitality, Tourism and Leisure, 13(1), 213–
220. https://doi.org/10.46222/ajhtl.19770720.500

[6] Bartaloni, Cecilia & ALDERIGHI, Marco. (2024). Are customers ready to be smart?
designing smart hotel experiences. Tourism and hospitality management. 30. 195-207.
10.20867/thm.30.2.4.

[7] Hutabarat, D. P., Saleh, R., Sebastian, L., Jonathan, J., & Arosha, S. M. N. (2023).
Smart system for controlling and monitoring electronic facilities in budget hotel rooms.
Indonesian Journal of Electrical Engineering and Computer Science, 30(2), 770–777.
https://doi.org/10.11591/ijeecs.v30.i2.pp770-777

[8] Liu, J., Hall, C. M., Zhu, C., & Ting Pong Cheng, V. (2023). Redefining the concept of
smart tourism in tourism and hospitality. Anatolia, 35(3), 566–578.
https://doi.org/10.1080/13032917.2023.2282712

[9] Ionescu, A.-M., & Sârbu, F. A. (2024). Exploring the impact of smart technologies on
the tourism industry. Sustainability, 16(8), 3318. https://doi.org/10.3390/su16083318

[10] Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of
intrinsic motivation, social development, and well-being. American Psychologist, 55(1),
68–78. https://doi.org/10.1037/0003-066X.55.1.68

[11] Melville, N. P. (2010). Information systems innovation for environmental

sustainability. MIS Quarterly, 34(1), 1–21. https://doi.org/10.2307/20721412

[12] Tussyadiah, I. (2020). A review of research into automation in tourism: Launching

the Annals of Tourism Research curated collection on artificial intelligence and robotics in
tourism. Annals of Tourism Research, 81, 102883.
https://doi.org/10.1016/j.annals.2020.102883

[13] Wiederhold, M., & Martinez, L. F. (2018). AI-driven energy efficiency in hospitality:
Evidence from smart hotels. Journal of Sustainable Tourism, 26(5), 759–774.
https://doi.org/10.1080/09669582.2017.1404611

[14] Tuomi, A., Tussyadiah, I. P., & Stienmetz, J. (2021). Applications and implications of
service robots in hospitality. *Cornell Hospitality Quarterly, 62*(2), 232–247.
https://doi.org/10.1177/1938965520923961

Pag. 254 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[15] Li, J., Bonn, M. A., & Ye, B. H. (2022). Hotel guest satisfaction during COVID-19
outbreak: The moderating role of crisis response strategy. Tourism Management, 93,
104618. https://doi.org/10.1016/j.tourman.2022.104618

[16] Venkatesh, V., Thong, J. Y. L., & Xu, X. (2012). Consumer acceptance and use of
information technology: Extending the unified theory of acceptance and use of
technology. MIS Quarterly, 36(1), 157–178. https://doi.org/10.2307/41410412

[17] Allison, P. D. (2001). *Missing data*. Sage Publications.

https://doi.org/10.4135/9781412985079

[18] Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). Sage.

[19] Guttentag, D. A. (2019). Progress on Airbnb: A literature review. Journal of

Hospitality and Tourism Technology, 10(4), 511–528. https://doi.org/10.1108/JHTT-08-
2018-0075

[20] Du, H., Li, J., So, K. K. F., & King, C. (2024). Artificial intelligence in hospitality
services: Examining consumers’ receptivity to unmanned smart hotels. Journal of
Hospitality and Tourism Insights. Advance online publication.
https://doi.org/10.1108/JHTI-06-2024-0548

Bibliography

Allison, P. D. (2001). *Missing data*. Sage Publications.

https://doi.org/10.4135/9781412985079

Alonso-Almeida, M. D. M., Bagur-Femenias, L., Llach, J., & Perramon, J. (2017).

Sustainability in small tourist businesses: The link between initiatives and performance.
Current Issues in Tourism, 20(7), 716–732.
https://doi.org/10.1080/13683500.2016.1201069

Arana-Landín, G., Uriarte-Gallastegi, N., Landeta-Manzano, B., & Laskurain-Iturbe, I.

(2023). The contribution of lean management—Industry 4.0 technologies to improving
energy efficiency. Energies, 16(5), 2124. https://doi.org/10.3390/en16052124

Baber, King, & Arsalan, Huzaifa. (2024). Maximizing hospitality industry efficiency: AI
applications for sustainability and customer satisfaction. ResearchGate.
https://doi.org/10.13140/RG.2.2.30819.98080

Pag. 255 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Bartaloni, Cecilia & ALDERIGHI, Marco. (2024). Are customers ready to be smart?
designing smart hotel experiences. Tourism and hospitality management. 30. 195-207.
10.20867/thm.30.2.4.

Chan, E. S., Okumus, F., & Chan, W. (2018). Barriers to environmental technology
adoption in hotels. Journal of Hospitality & Tourism Research, 42(5), 829–852.
https://doi.org/10.1177/1096348018777046

Çeltek, Evrim. (2023). Analysis of Smart Technologies Used in Smart Hotels. Journal of
Business Research - Turk. 10.20491/isarder.2023.1754.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008

Dianawati, N., Saepudin, P., Misran, M., Sinaga, E., Putra, F., & Susanto, E. (2024). The
impact of smart hotel technology on guest satisfaction and loyalty: A user competency
perspective. African Journal of Hospitality, Tourism and Leisure, 13(1), 213–220.
https://doi.org/10.46222/ajhtl.19770720.500

Du, H., Li, J., So, K. K. F., & King, C. (2024). Artificial intelligence in hospitality
services: Examining consumers’ receptivity to unmanned smart hotels. Journal of
Hospitality and Tourism Insights. Advance online publication.
https://doi.org/10.1108/JHTI-06-2024-0548

Dykins, R. (2023). Leyeju smart hotels offer entirely touch-free stays in China.
Globetrender. https://globetrender.com/2020/08/24/leyeju-smart-hotels-touch-free-stays-
china/

Elliot, S. (2016). Green IS and consumer behavior: A framework for research. Information
Systems Journal, 26(3), 183–206. https://doi.org/10.1111/isj.12093

Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). Sage.

Gajić, T., Petrović, M. D., Pešić, A. M., Conić, M., & Gligorijević, N. (2024). Innovative
approaches in hotel management: Integrating artificial intelligence (AI) and the Internet of
Things (IoT) to enhance operational efficiency and sustainability. Sustainability, 16(17),
7279. https://doi.org/10.3390/su16177279

Guttentag, D. A. (2019). Progress on Airbnb: A literature review. Journal of Hospitality

and Tourism Technology, 10(4), 511–528. https://doi.org/10.1108/JHTT-08-2018-0075

Hutabarat, D. P., Saleh, R., Sebastian, L., Jonathan, J., & Arosha, S. M. N. (2023). Smart
system for controlling and monitoring electronic facilities in budget hotel rooms.
Indonesian Journal of Electrical Engineering and Computer Science, 30(2), 770–777.
https://doi.org/10.11591/ijeecs.v30.i2.pp770-777

Pag. 256 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Ionescu, A.-M., & Sârbu, F. A. (2024). Exploring the impact of smart technologies on the
tourism industry. Sustainability, 16(8), 3318. https://doi.org/10.3390/su16083318

Jones, P., Hillier, D., & Comfort, D. (2020). Sustainability in the hospitality industry:
Some personal reflections on corporate challenges and research agendas. International
Journal of Contemporary Hospitality Management, 32(1), 53–67.
https://doi.org/10.1108/IJCHM-02-2019-0160

King, B., & Arsalan, H. (2024). Maximizing hospitality industry efficiency: AI

applications for sustainability and customer satisfaction. ResearchGate.
https://doi.org/10.13140/RG.2.2.30819.98080

Li, J., Bonn, M. A., & Ye, B. H. (2022). Hotel guest satisfaction during COVID-19
outbreak: The moderating role of crisis response strategy. Tourism Management, 93,
104618. https://doi.org/10.1016/j.tourman.2022.104618

Liu, J., Hall, C. M., Zhu, C., & Ting Pong Cheng, V. (2023). Redefining the concept of
smart tourism in tourism and hospitality. Anatolia, 35(3), 566–578.
https://doi.org/10.1080/13032917.2023.2282712

Makivić, R., Vukolić, D., Veljović, S., Bolesnikov, M., Dávid, L. D., Ivanišević, A., Silić,
M., & Gajić, T. (2024). AI impact on hotel guest satisfaction via tailor-made services: A
case study of Serbia and Hungary. Information, 15(11), 700.
https://doi.org/10.3390/info15110700

Melville, N. P. (2010). Information systems innovation for environmental sustainability.

MIS Quarterly, 34(1), 1–21. https://doi.org/10.2307/20721412

M'hamed, E. G., & Idrissi, R. E. (2024). Transforming hospitality: Harnessing artificial

intelligence for enhanced guest experience and operational efficiency. In Y. Farhaoui
(Ed.), Artificial intelligence, big data, IoT and block chain in healthcare: From concepts to
applications (pp. 1–15). Springer. https://doi.org/10.1007/978-3-031-65014-7_17

Patrichi, I. C. (2024). AI solutions for sustainable tourism management: A comprehensive

review. Journal of Information Systems & Operations Management, 18(1), 172–185.

Pereira, V., Silva, G. M., & Dias, Á. (2021). Sustainability practices in hospitality: Case
study of a luxury hotel in Arrábida Natural Park. Sustainability, 13(6), 3164.
https://doi.org/10.3390/su13063164

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of
intrinsic motivation, social development, and well-being. American Psychologist, 55(1),
68–78. https://doi.org/10.1037/0003-066X.55.1.68

Pag. 257 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Sarode, A. V., & Lukose, W. (2025). Evaluation of integration of artificial intelligence in

enhancing guest experiences. Journal of Hospitality and Tourism Technology. Advance
online publication.

Smith, H. J., Dinev, T., & Xu, H. (2011). Information privacy research: An
interdisciplinary review. MIS Quarterly, 35(4), 989–1016.
https://doi.org/10.2307/41409971

Tuomi, A., Tussyadiah, I. P., & Stienmetz, J. (2021). Applications and implications of
service robots in hospitality. *Cornell Hospitality Quarterly, 62*(2), 232–247.
https://doi.org/10.1177/1938965520923961

Tussyadiah, I. (2020). A review of research into automation in tourism: Launching the

Annals of Tourism Research curated collection on artificial intelligence and robotics in
tourism. Annals of Tourism Research, 81, 102883.
https://doi.org/10.1016/j.annals.2020.102883

Venkatesh, V., Thong, J. Y. L., & Xu, X. (2012). Consumer acceptance and use of
information technology: Extending the unified theory of acceptance and use of
technology. MIS Quarterly, 36(1), 157–178. https://doi.org/10.2307/41410412

Wiederhold, M., & Martinez, L. F. (2018). AI-driven energy efficiency in hospitality:

Evidence from smart hotels. Journal of Sustainable Tourism, 26(5), 759–774.
https://doi.org/10.1080/09669582.2017.1404611

Wirtz, J., Patterson, P. G., Kunz, W. H., Gruber, T., Lu, V. N., Paluch, S., & Martins, A.
(2018). Brave new world: Service robots in the frontline. Journal of Service Management,
29(5), 907–931. https://doi.org/10.1108/JOSM-04-2018-0119

Zhang, Q., Khan, S., Khan, S. U., Khan, I. U., & Mehmood, S. (2024). Tourist
Motivations to Adopt Sustainable Smart Hospitality: An Innovation Resistance Theory
Perspective. Sustainability, 16(13), 5598. https://doi.org/10.3390/su16135598

Pag. 258 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ARTIFICIAL SOCIAL INTELLIGENCE AND THE TRANSFORMATION OF

HUMAN INTERACTION BY ARTIFICIAL INTELLIGENCE AGENTS

Alexandru PÎRJAN1
Dana-Mihaela PETROŞANU2

Abstract
Artificial Intelligence (AI) has profoundly influenced numerous sectors, with Artificial
Social Intelligence (ASI) emerging as an extremely important area focusing on AI agents'
ability to understand, interpret, and engage within human social contexts. This article
provides an analysis of ASI and its changing impact on human interaction. We trace the
evolution of ASI from early concepts to complex computational agents capable of nuanced
social behaviors, highlighting the interdisciplinary foundations spanning computer science,
psychology, sociology, and ethics. The paper makes an in-depth analysis into the theoretical
basis and into the key technologies that enable ASI. We explore the emergence of AI agents
in diverse social settings and analyze the mechanisms through which they are reshaping
communication dynamics, group interactions, social norms, empathy, and concepts of
identity and authenticity. Furthermore, the article presents applications and case studies
across very important domains, analyzing both the potential benefits and inherent
challenges. A significant portion is dedicated to the ethical, legal, and societal implications
(ELSI), addressing concerns related to transparency, privacy, bias, accountability, trust, and
psychological well-being. We identify key challenges and open research questions and
discuss future directions, exploring enhanced technologies, the pursuit of generalized social
intelligence, and the potential for human-AI symbiosis. The conclusion emphasizes the
need for a collaborative, ethically grounded, and systemic approach to guide the
development and deployment of ASI, ensuring it serves to augment human flourishing and
enrich the collective social experience rather than diminishing it.
Keywords: Artificial Social Intelligence, Computational Agents, Artificial Intelligence, AI
Agents, Human Social Contexts, Human Interaction, Ethical Legal Societal Implications,
Challenges and Open Research Questions
JEL Classification: O3, O33, O34, O35, O36, O38

1. Introduction

1PhD Hab. Full Professor, School of Computer Science for Business Management, Romanian-American
University, alexandru.pirjan@rau.ro, corresponding author
2PhD Lecturer, Department of Mathematics-Informatics, National University of Science and Technology

Politehnica Bucharest, Romania, dana.petrosanu@upb.ro

Pag. 259 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Research and development in the field of AI have experienced unprecedented growth over
the last several decades, influencing numerous sectors of society. Economic, political, and
cultural domains have all felt the impact of increasingly complex AI systems. Nevertheless,
within the emergence of AI in various industries, ranging from finance [1] and healthcare
[2] to entertainment [3] and transportation [4], one of the most important and impactful
aspects has been the rise of ASI. This area focuses on the ability of AI systems to control,
interpret, and engage in social contexts, thereby influencing the way in which humans
interact with each other and with computational agents [5]. An emphasis on social
capabilities compels AI research to draw on theories and practices from psychology [6],
sociology [7], linguistics [8], neuroscience [9] and cognitive science [10], among other
fields, highlighting that ASI requires an inherently interdisciplinary approach.
In its most basic sense, ASI aims to replicate or approximate various dimensions of human
social behaviors and interactions within artificial agents. Rather than simply performing
computations or providing information, socially intelligent AI agents are designed to detect
emotional expressions [11], infer intentions [12], respond empathically [13], and adapt to
changing social environments [14]. While early AI research focused heavily on problem-
solving, logic, and symbolic reasoning, the next generation of AI systems has shifted
attention to nuanced interpersonal dynamics [15], complex communication patterns [16],
and adaptive social behaviors [17]. This transformation is fueled by the growing recognition
that human interaction encompasses much more than the exchange of factual content. It
involves shared context, subtle emotional cues, cultural norms, relational history, and a
wide range of evolving social meanings.
The significance of ASI is becoming more pronounced as AI technologies begin to mediate
human communication on a large scale. Online platforms, messaging services, virtual
assistants, and social media channels increasingly rely on AI algorithms that filter
information, recommend content, and simulate social presence. Advancements in Natural
Language Processing (NLP) [18], Machine Learning (ML) [19], and computer vision [20]
have enabled the creation of conversational agents [21], social robots [22], and other AI-
driven systems that can carry out tasks in ways that feel surprisingly personable and
context-aware [23]. These developments prompt a recalibration of human interaction, as AI
moves from an automated background function to a socially interactive presence capable
of influencing interpersonal dynamics.
At the core of this progression is the fundamental question of how humans respond to, adapt
to, and co-evolve with AI entities that exhibit social intelligence. The phenomenon
transforms the user experience into technology-mediated settings, along with the broader
landscape of social norms, ethical considerations, psychological constructs of identity and
agency. It is therefore very important to investigate the interaction between AI systems with
social capacities and the human communities that adopt them [24]. The transformation has
potential to yield positive outcomes such as enhanced accessibility, improved social
Pag. 260 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

connectivity, and the facilitation of collaborative tasks across cultural and geographic
distances [19]. Nevertheless, it also raises critical challenges, including privacy concerns
[25], bias in AI decision-making [23], the erosion of certain human to human interaction
skills [26], and a reevaluation of how authenticity is defined and maintained in the digital
and physical spaces [27].
Within this broader context, the article provides an in-depth exploration of ASI and of the
way in which it supports and drives a transformation of the human interaction. The
introduction focuses on the historical roots, theoretical frameworks, technological
advancements, ethical considerations, and future directions that comprise this complex
topic. By examining the motivations, scope, and significance of developing socially aware
AI agents, it becomes possible to put forward the broader implications for human society,
as well as the responsibilities of designers, policymakers, and researchers involved in this
cutting-edge field. The following sections describe the motivations behind the field's rapid
expansion, the objectives that guide ongoing research, the significance of interaction
between human beings and AI technology for social experiences.
The evolution of AI, from symbolic reasoning to data-centric methods, has led to
breakthroughs in image recognition [28], NLP [29] and robotics [30]. Nevertheless, many
systems lack social awareness and interaction. The development of socially aware AI aims
to replicate human social intelligence, including emotional recognition and perspective-
taking, to enable more natural and ethical human-machine collaborations [31].
A particular interest consists in studying the field of socially aware AI, examining its
theoretical foundations, technological innovations, historical developments, and ethical
implications. We have analyzed how AI-based social intelligence differs from classical AI,
evaluate key metrics for assessing AI performance in social contexts, and highlight the
technological components enabling socially intelligent AI. The article has also analyzed
current research within a historical perspective, discussing the impact of AI on human
interaction by presenting applications and case studies.
The socially intelligent AI marks a significant shift in human-AI interaction [31], moving
beyond traditional tool-like roles to more interactive and considerable positions. ASI
introduces new forms of communication and collaboration [15], raising questions about
trustworthiness [32], authenticity [27] and power dynamics in digital spaces [33]. This
evolution impacts technology acceptance [34], community formation [35], and societal
norms [36], prompting philosophical reflections on personhood [37] and human identity
[31].
The present article is organized to provide a comprehensive examination of ASI,
culminating in a full perspective that brings together various aspects. Therefore, following
this introduction, the second section of the article makes an in-depth analysis into the
foundations of ASI, the third one studies the theoretical and technological underlying

Pag. 261 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

elements, the fourth section analyses the emergence and evolution of AI agents in human
social contexts, the fifth one makes an in-depth analysis into the transformation of human
interaction, the sixth section highlights applications and case studies, being followed by the
seventh one that analyses the ethical, legal, and societal implications, the eighth section that
investigates challenges and open research questions, while the ninth one analyses future
directions and opportunities, along with a detailed conclusions section.

2. Foundations of ASI
ASI is a rapidly evolving domain that seeks to provide computational systems with
capabilities that go far beyond simple data processing or pattern recognition. The emphasis
lies in designing AI agents that can traverse social contexts with sensitivity, adaptability,
and awareness of both the emotional states and the behavioral norms of the people around
them. This field addresses the question of how machines can acquire and exhibit behaviors
that typically demand high levels of human-like empathy, context understanding, and
interpersonal intuition. Over the years, there has been growing recognition of the necessity
for AI to perform tasks efficiently and to be able to engage in dynamic, nuanced interactions
that reflect a certain level of social complexity [2,36].
The present section makes an in-depth analysis into the foundational aspects of ASI by first
clarifying what "social intelligence" means in the context of AI research. It then puts
together AI social skills with more conventional AI approaches, highlighting what sets
socially capable AI apart from standard ML or knowledge-based systems. Following this,
the cognitive and psychological aspects that inspire how AI agents can model and respond
to human behavior has been discussed, with an emphasis on both classical and
contemporary insights from psychology and neuroscience. Finally, key metrics and
benchmarks for evaluating AI's social intelligence are analyzed with a view to how the field
measures progress and determines the extent to which AI agents succeed in meaningful
interactions between humans and AI technology.
Social intelligence in AI aims to give machines the ability to comprehend communication,
predict human behavior, and act appropriately. This involves more than just language
proficiency, it requires competencies like reading emotions and reflecting empathy. The
field, at the intersection of computer science, behavioral psychology, and interaction design,
has evolved from simple chatbots to systems that consider multimodal communication and
cultural context [38].
Conventional AI excels at well-defined tasks with labeled datasets or explicit mathematical
models, such as image classification, speech recognition, or strategic decision-making in
deterministic environments [39]. Nevertheless, social intelligence demands that AI agents
process, classify, interpret, and respond to subtle human interaction complexities, including
emotional nuances, cultural specificities, and historical patterns of behavior [14,36,38].
Pag. 262 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

A socially intelligent AI must integrate empathy [5,13], trust-building [40], politeness [41],
cooperation [42], or conflict resolution [43] into its decision-making process [33],
considering not just immediate outcomes but also their impact on future relationships,
group norms, and individual well-being. This change in perspective fundamentally alters
AI model design, training, and evaluation.
Context is very important in the field of AI, especially in social interactions. Image
classification models focus on visual features, while language models consider textual
syntax and semantics within a domain [44]. In contrast, social interactions are dynamic,
encompassing real-time emotions, past interactions, environmental cues, and cultural norms
[45]. For instance, scheduling a meeting differs from persuading someone to attend,
requiring social skills like communication and understanding group dynamics.
Conventional AI approaches that treat communication as a simple input-output problem
may fail in these nuanced contexts due to the lack of relational and emotive dimensions.
Evaluation metrics for socially intelligent AI differ from those used in typical AI systems.
Accuracy, precision, recall, and F1 scores do not capture the richness of social interactions
[46]. These metrics provide insights for well-defined tasks but overlook aspects like natural
responses [47], trust advancement [32,40], or conflict resolution [43]. Users might value
perceived authenticity or comfort in social settings over computational accuracy. An AI
system could be considered socially skilled for understanding implicit social rules, even
with occasional errors, while conventional AI systems usually face harsh judgment for
small inaccuracies in classification, regardless of social rapport.
The initiative to develop socially skilled AI aims to make machines integrated participants
in our social and organizational landscapes, rather than isolated tools. This requires
rethinking system architectures [48], training paradigms [49], and design philosophies [50].
Researchers focus on interdisciplinary knowledge, including social psychology,
communication studies, organizational behavior, and philosophical inquiries into
intelligence, in order to create AI agents that can authentically engage with human social
realities [14]. These agents handle the cognitive load of processing high-dimensional data
and the emotional and relational dimensions of human communities. While conventional
AI is important for technical tasks, socially intelligent AI systems carry additional
complexity and responsibilities in their operations and interactions with human
stakeholders.
The design of socially intelligent AI is informed by cognitive, behavioral, and neuroscience
research [9,10,17]. Psychologists study how individuals perceive, interpret, and respond to
others' emotions, intentions, and beliefs, focusing on uncovering details regarding empathy,
theory of mind, and social reasoning. Computationally, researchers model these processes
within an AI architecture, drawing parallels between human cognitive functions and
computational modules for perception, memory, attention, and decision-making processes.

Pag. 263 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

In contrast to humans who rely on neural mechanisms, socially intelligent AI systems use
algorithmic and data-driven methods to replicate or approximate these capabilities [51].
In classic cognitive science, theory of mind refers to the ability to attribute mental states to
oneself and others, enabling prediction of behavior and adaptation of actions [52]. AI
systems aiming for social intelligence need theory of mind to anticipate user or group
responses [53]. Computationally, this can be approached using probabilistic models
tracking latent states representing others' thoughts or feelings. Over time, the AI updates
assumptions based on the observed behaviors and refines their predictions, putting the basis
for socially aware planning and collaboration, allowing AI agents to adapt strategies when
the goals shift [31].
Empathy, the ability to understand and experience another's emotions, enables support,
compassion, and meaningful relationships [5,13]. Capturing emotional signals from text,
voice, or facial expressions and mapping them to internal representations guides empathic
AI responses. Early attempts focus on keyword detection or sentiment analysis, while
contemporary approaches use deep learning for nuanced emotional state detection [11].
Some systems integrate physiological data or contextual variables. The challenge consists
in ensuring that empathic responses go beyond superficial imitations. AI should understand
narrative context, user background, and social norms in specific cultures or situations.
Beyond empathy and theory of mind, other cognitive capacities like attention and memory
significantly influence social interactions. Humans selectively focus attention based on
popularity, relevance, and social cues to manage group settings. AI systems with social
intelligence need efficient mechanisms to direct computational resources to important
stimuli, like dynamic attention models that shift focus based on conversation topics,
emotional tone, or new individuals [54]. Memory systems must account for long-term
language or cultural knowledge and short-term contextual details. This integrated approach
to attention and memory creates fluid, contextually grounded interactions. Reinforcement
learning from human feedback forms the basis for adaptation and improvement. Social
interactions involve trial and error, and AI can benefit from interactive learning loops that
capture user feedback, either explicit or implicit. As AI refines its models, it can develop
more accurate expectations about actions that encourage cooperation, trust, or positive
emotions. The convergence of empathy, theory of mind, attention, memory, and adaptive
learning provides a framework for AI researchers to build systems that mirror human social
interaction [5,13,52].
Developers must be mindful of the potential pitfalls of anthropomorphizing AI. Humans
tend to attribute agencies and intention to entities exhibiting social cues, even if they are
algorithmically generated. This inclination can be exploited for engaging user experiences,
but it raises ethical dilemmas about deception and emotional reliance on AI companionship.
Understanding how humans process social signals is very important for crafting
responsible, respectful AI systems that are transparent about their computational nature.
Pag. 264 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Aligning AI design with these cognitive insights ensures social intelligence coexists
harmoniously with human emotions and psychology [38] .
As AI research advances, there is a growing need for robust metrics and benchmarks to
quantify social capabilities. Traditional AI evaluations like accuracy, precision, and recall
are useful for tasks like object recognition or language translation, but they fail to capture
the depth of social interactions. Researchers must translate subjective human judgments
about empathy, rapport, trust, and adaptability into metrics that guide algorithmic
development and compare systems [32].
One approach involves using standardized role-playing scenarios or simulated
environments where human participants interact with AI agents under controlled
conditions. Researchers can measure variables like user satisfaction, perceived empathy,
willingness to disclose personal information, or physiological signals to assess emotional
impact [55]. Surveys and questionnaires collect subjective evaluations of the AI's
performance, capturing elements like perceived social presence, trustworthiness, and
attentiveness. While subjective measures vary, aggregating results provides insights into
how well AI meets social expectations. As scenarios become more immersive, new
opportunities arise to evaluate AI's capacity for nonverbal communication and spatial
interaction.
Objective measures assess social intelligence. Turn-taking fidelity quantifies how an AI
adapts to conversational rhythms. Metrics assess topical coherence and topic shifts without
losing user engagement. Some systems use sentiment analysis or emotion detection to
measure alignment between user and AI responses [56]. Consistent supportive responses
indicate empathic alignment, but quantifying it is an extremely complex task. Blending
objective data with subjective user feedback creates multi-dimensional performance
indicators [38].
Benchmark datasets standardize evaluations. Large-scale conversation datasets test AI's
ability to generate context-appropriate, empathetic, or persuasive responses. Specialized
elements contain emotionally charged dialogues, negotiations, or collaborative tasks
requiring strong social components. Systems are evaluated against these benchmarks to
measure response alignment with human examples or social criteria. Nevertheless, a
universal benchmark is elusive due to varying social norms, cultural references, and
contextual cues. Efforts develop more culturally diverse benchmarks, but inherent
variability makes any single resource incomplete [45,57].
Progress in AI social competence depends on external validation through competitions and
challenges organized by academic conferences and research consortia. These competitions
simulate complex social interactions, assessing multi-party negotiation, collaborative
storytelling, and group consensus-building. Successful systems demonstrate linguistic
skills and effective interpersonal dynamics. These competitive settings promote community

Pag. 265 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

convergence on best practices and refined metrics, leading to standardized evaluation

protocols for ASI. However, the diversity of real-world social contexts emphasizes the need
for adaptable AI social skills [58].
Iterative metric refinement extends beyond academic research. Standardized evaluations
assess conversational agents, social robots, and AI-driven platforms for customer-facing
tasks [22,23]. Regulatory bodies and stakeholders ensure compliance with transparency,
safety, and fairness standards. A consensus may emerge on key indicators of social
proficiency, such as emotional engagement, rapport continuity, privacy respect, and cultural
sensitivity. Until then, the domain must balance objective and subjective metrics, universal
standardization, and localized contextualization to measure AI's transition into genuinely
social domains.
Social intelligence in AI combines ideas from computer science, cognitive psychology,
neuroscience, sociology, and ethics. It reconciles human social competency with AI
algorithmic methods. Distinguishing socially intelligent AI from traditional AI highlights
its complexity and need for context-awareness, emotional resonance, and interpersonal
nuance. Cognitive and psychological aspects like theory of mind, empathy, attention, and
memory support AI architectures that mimic human interaction. In addition, measuring
social intelligence in AI requires novel metrics beyond accuracy and efficiency, considering
richness and sensitivity in genuine interpersonal interactions.
Building an AI system that truly integrates into human social fabrics goes beyond achieving
high scores on narrowly defined tasks. It involves understanding how human relationships,
cultural norms, emotional expressions, and interactive protocols influence community
interactions. When AI agents learn dynamically, show empathy, respect social norms, and
adapt, they approach true social intelligence [14,38]. This vision envisions AI entities
seamlessly collaborating with people in various settings, enhancing the quality of human
experience rather than displacing or disrupting it. One must take into account that this vision
requires refining ethical, legal, and societal frameworks for socially intelligent AI [19]. The
next sections explore the theoretical and technological underlying elements of such AI
systems, their emergence, and their impact on human communities.

3. Theoretical and Technological Underlying Elements

This section focuses on the core theoretical and technological foundations that make ASI a
reality. The interaction among ML, NLP, computer vision, reinforcement learning, and
multi-agent systems provides numerous methods and frameworks for creating AI agents
capable of understanding social cues, facilitating human-like communication, and adapting
to dynamic group environments. Socially aware AI have to base upon fundamental
principles from cognitive science and psychology, because the design of these agents
cannot merely rely on standard computational approaches [6,10]. The more one advances
Pag. 266 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

in this direction, the clearer it becomes that effective social intelligence arises from the tight
integration of algorithmic innovations with nuanced models of human interaction. These
interactions involve diverse signals in the forms of language, emotion, and behavioral cues,
all of which shape how AI agents perceive and respond within human environments.
The strategic significance of these aspects can be understood considering the evolving
applications of AI across fields such as healthcare, education, corporate environments, and
the public sector. A robust understanding of the algorithms and theories is important for
steering the ongoing transformation of the interaction between humans and AI technologies.
By providing more details on the mechanisms by which AI systems learn, interpret, and
modify social behaviors, we obtain deeper insights into how to design AI that operates with
empathy and accountability [57]. Technical breakthroughs in these domains will redefine
how we conceptualize interaction in the future, from small-scale personal companionship
solutions to large-scale policymaking and governance tools.
ML is the backbone of socially aware AI, enabling models to capture input-output mappings
and contextual variables of human interactions [19]. While traditional supervised learning
approaches can be adapted to label social cues, unsupervised and semi-supervised methods
are very important for uncovering latent social patterns. Despite advancements in
architecture like multi-modal transformer models and generative modelling, addressing
biases and ensuring ethical deployment remain ongoing challenges [58].
NLP occupies a central role in ASI, enabling conversational agents to interpret user input
and generate coherent responses. Large pre-trained language models, refined for social
contexts, require additional training data and ethical guidelines to ensure polite and
empathetic interactions. Future advancements in NLP aim to incorporate extralinguistic
context, emotional intelligence, and potentially gesture recognition, creating more
sophisticated and empathetic conversational agents [29,46,56] .
Computer vision is very important for AI systems in order to understand human
communication, particularly through emotion and gesture recognition. While traditional
methods rely on labelled datasets, real-world scenarios demand more nuanced approaches,
including temporal modelling and robust performance across diverse conditions.
Combining emotion and gesture recognition with contextual cues, such as scene
understanding and multimodal fusion, enables more accurate inference of complex
emotional states, creating the premises for truly socially intelligent AI systems [20].
Reinforcement Learning (RL) is a powerful framework for training socially intelligent AI
agents in order to optimize sequential decisions in uncertain environments. RL enables
agents to adapt their behavior based on rewards capturing socially desirable outcomes, such
as user satisfaction and group consensus. One must consider that designing complex reward
functions, handling partial observability, and ensuring scalability and ethical behavior
remain significant challenges [55].

Pag. 267 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Multi-agent systems, combining ML, NLP, computer vision, and reinforcement learning,
enable complex social interactions between AI agents and humans [59]. These systems,
which can exhibit collaborative intelligence, require trust and reputation mechanisms,
effective communication protocols, and game theory-based coordination strategies. As
multi-agent systems become more prevalent in everyday life, rigorous theoretical
foundations and robust technological implementations are very important to ensure they
align with human values and social norms.

4. Emergence and Evolution of AI Agents in Human Social Contexts

Ever since the earliest days of computing, researchers and thinkers have envisioned whether
machines could emulate, simulate, or augment human social interaction. This interest traces
back to post-World War II efforts in computer science, where pioneers like Alan Turing
questioned whether machines could exhibit behaviors indistinguishable from humans.
Turing's famous test for machine intelligence, introduced in 1950 [60], invited a wide range
of ethical, philosophical, and technical inquiries, many of which were directly linked to
how a machine's language use could approximate or surpass human capacities in
conversation. During those formative decades, there was less public focus on the social
dimension of these interactions. The perception about AI at that time concentrated more
acutely on reasoning, problem-solving, and symbolic logic. Attempts to explore the social
potential of AI can be found in even the earliest attempts at building programs capable of
dialogue [60].
The mid-1960s saw a notable departure from purely logic-based or mathematically oriented
AI when Joseph Weizenbaum introduced ELIZA, a program that engaged in rudimentary
conversation by rephrasing user inputs in the style of a Rogerian psychotherapist.
Weizenbaum's intentions were neither to create a companion nor to replicate human
empathy [60]. However, the way people reacted to ELIZA revealed something profound,
namely despite the system's obvious limitations, many users attributed human-like qualities
and emotional understanding to the program. This phenomenon of anthropomorphizing
what was in essence a simple pattern-matching script, signaled that the social dimensions
of AI could not be dismissed as an afterthought. Even then, the media latched onto the idea
of speaking machines, which intrigued the broader public and created an enthusiasm for
imagining a future where computers might be embedded in human social contexts [60].
In the following decades, from the 1970s through the 1990s, AI witnessed "several winters"
where enthusiasm and funding took a downturn. These periods were often tied to a
mismatch between grand promises and the technical realities of hardware constraints,
software complexity, and the challenges of generalization beyond narrow, rule-based
problem domains. Nevertheless, research in areas such as knowledge representation, expert
systems, and Natural Language Understanding (NLU) continued to progress. While the

Pag. 268 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

focus was rarely placed on "social intelligence" directly, these advances created the
foundational architecture that would later allow for more complex, context-aware, and
interactive AI applications [46]. Systems like SHRDLU, which demonstrated the ability to
interpret and act upon written commands about a virtual blocks world, hinted at the potential
of AI to engage in conversation about shared contexts [57]. Even though these dialogues
were strictly constrained, they represented a step forward in generating responses that took
account of user input in a dynamically evolving situation [57].
In parallel, robotics labs around the world began to explore more physically embodied
forms of AI. Machines that could move, sense their environment, and adapt to unpredictable
contexts gave rise to early social robotics, though such projects were generally surpassed
by more specialized or industrial applications. The notion of physical embodiment would
later prove to be an important factor in how society received AI agents, as robots that could
manage human spaces and physically assist people, aspects that generated enthusiasm,
curiosity, and sometimes fear. By the late 1990s, humanoid robots started to appear in
academic settings, often used for research on gait, balance, and automated movement. Some
of these projects, while primarily technical, indicated the future integration of robots into
domestic and care environments, signaling a transition from purely operational tasks to
those that demanded social awareness [31].
As the 20th century gave way to the 21st, the internet and mobile devices revolutionized
how people communicated, stirring fresh interest in socially oriented AI. Digital
infrastructure evolved rapidly, with high-speed connectivity enabling large-scale data
collection from user interactions. These data-rich environments became suitable for ML
approaches, opening the door for systems that could learn from static databases and also
from continuous streams of human behavior in real-world contexts. By this point, a
historical pattern was beginning to manifest progress in AI's social capabilities followed
closely behind by innovations in computational power, data availability, and fundamental
algorithmic improvements. While the earliest conceptual explorations might have been
rooted in philosophical questions about conversation, by the early 2000s, it was clear that
the social dimension of AI was not a marginal or speculative curiosity. It was moving into
the mainstream of research and, increasingly, into public awareness [25].
Interactive AI has evolved from early chatterbots like ELIZA to sophisticated systems
capable of NLP [56,61] and generation [62]. Advancements in neural networks [63],
computer vision [20], and robotics [30] have enabled AI to interpret visual signals and
engage in social interactions [45,64]. The widespread adoption of deep learning techniques
in the early 2010s further improved AI capabilities, leading to the development of more
advanced chatbots and virtual assistants [65].
Advancements in AI have led to the development of advanced social robots, chatbots, and
virtual assistants [46]. These agents, equipped with advanced sensors and computational
frameworks, can interpret and respond to human cues, engage in dynamic dialogues, and
Pag. 269 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

adapt to user emotions and contexts. The integration of large language models and multi-
modal interaction further enhances their capabilities, enabling them to simulate social
presence and seamlessly blend into daily life.
Public perception of AI agents in social contexts is shaped by cultural accounts, media
portrayals, and historical experiences. While some regions, like Japan and South Korea,
embrace AI due to positive cultural perceptions, others are concerned about job
displacement, privacy, and ethical implications. The acceptance and integration of AI into
society vary widely across demographics, nations, and historical contexts, influenced by
factors like trust, cultural values, and technological advancements.

5. Transformation of Human Interaction

The advent of ASI has created an extremely important change in human activities, in which
conversations, social gatherings, organizational meetings, and even personal relationships
are being fundamentally changed by the presence of AI agents [11–14] . These agents, with
capabilities ranging from NLP to emotion recognition, assist human beings in completing
tasks and also actively participate in communicative settings that have historically been the
exclusive domain of human-to-human interaction. Such participation leads to a profound
reconfiguration of the ways in which people convey information, interpret social cues, form
judgments, and develop empathy. Over time, the integration of AI agents into everyday
communication holds the potential to transform cultural values, ethical norms, and
conceptions of identity [15–17]. The following sections explore these developments by
examining the specific impacts of AI agents on communication dynamics, the mediation of
group interactions, the evolution of social norms and etiquette, the facilitation or hindrance
of human empathy, and the larger implications for personal identity and authenticity in
environments that blend human and AIs.
The integration of AI agents into communication channels is changing social and
organizational ecosystems. AI agents, with their contextual awareness and real-time
response capabilities, influence communication styles, pacing, and even non-verbal cues.
As individuals adapt to the AI's presence, new norms of "AI etiquette" emerge, blurring the
lines between human and machine input in communication.
The integration of AI agents into group settings, from small teams to large social networks,
is changing collaboration, conflict resolution, and resource sharing. AI mediators, often
acting as facilitators or leaders, can detect conflicts and propose data-driven solutions,
potentially leading to more impartial decisions. Nonetheless, this raises concerns about
autonomy, creativity, and the potential perpetuation of biases, highlighting the need for
transparency and oversight in AI-mediated group interactions [31,45] .
The integration of AI agents into personal and professional spheres is leading to the
emergence of new social norms and etiquette. These new conventions, which often develop
Pag. 270 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

naturally, address the challenge of interacting with machines that mimic social behaviors
[17]. As AI systems become more complex, etiquette rules will continue to evolve,
reflecting the hybrid nature of modern interactions and potentially leading to numerous
localized "AI etiquette" traditions [66].
The impact of AI on human empathy has a high degree of complexity. While AI can
facilitate empathetic engagement through tools for emotional analysis and conflict
resolution, it can also create dependencies and artificial simulations that weaken human-to-
human emotional bonds. The net impact of AI on empathy depends heavily on design
decisions, cultural interpretations, and user choices [5,13].
The introduction of artificial intelligence into social interactions prompts deep reflection on
concepts like human identity and authenticity. AI-driven agents, capable of producing
coherent and contextually relevant interactions, increasingly model our digital and social
environments. Such involvement significantly influences the way individuals form and
perceive identity across personal, community, and professional contexts, impacting factors
such as creativity, interpersonal networks, and individual autonomy. Consequently, the
distinction between identities constructed by humans and those influenced or created by AI
becomes progressively ambiguous.

6. Applications and Case Studies

ASI, capable of interpreting, responding to, and anticipating social cues, has rapidly moved
beyond theoretical constructions and laboratory experiments to become a significant force
in various human-centered domains. Social Artificial Intelligence (AI) agents are becoming
increasingly integrated into important societal sectors, including healthcare, education,
corporate settings, and governmental institutions, thereby changing the operational
practices, communication norms, and individual user experiences. The incorporation of
socially capable AI is evolving from an optional enhancement to an important component
for improving service quality, enhancing user satisfaction, and broadening access to
resources. Nevertheless, the global proliferation of these technologies is uneven, presenting
substantial ethical and structural complexities contingent upon variations in technological
infrastructure, regulatory directives, and established societal values. Despite these
impediments, an analysis of existent use cases and case studies highlights both the
significant potential, and the inherent challenges associated with deploying social AI across
diverse domains [62].
The functionality of socially aware AI relies upon several core technologies, namely
Machine Learning (ML) which enables personalization, Natural Language Processing
(NLP) that facilitates conversational capabilities, and Computer Vision which allows for
the recognition and interpretation of human behavior. Furthermore, Reinforcement
Learning techniques permit continuous adaptation, particularly within dynamic social
Pag. 271 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

environments. Successful deployment requires rigorous methodologies for data collection

and interpretation, along with ethically guided AI outputs. Domain-specific ethical
considerations are very important, healthcare and education need strict privacy and security
protocols due to the handling of sensitive personal data [30,40], while corporate and
governmental applications emphasize the importance of fairness, accountability, and
inclusivity [15].
The changing impact of AI is manifested in specific applications. Within healthcare, social
AI agents contribute to mental health therapeutic interventions and help mitigate
communication barriers [2,46,61]. In education, adaptive tutoring systems provide
personalized learning experiences and support the development of social competencies
[63]. Corporate environments leverage AI for enhanced coordination, although this raises
concurrent concerns regarding employee privacy and the potential devaluation of human
tasks [57]. In the situations of social media and personal digital life, AI-driven content
recommendation presents an ethical dilemma, balancing beneficial personalization against
the risk of promoting manipulative echo chambers [35]. Public administration applications
highlight AI's role in changing civic engagement, policy decisions, and citizen trust [57].
Exploring these domains reveals AI's profound impact on human connections, cooperation,
and agency in an automated world.
In the following there are depicted the applications of socially capable AI across five major
areas: healthcare, education, corporate environments, social media and personal life, along
with governance and public administration. There are also explored operational
mechanisms, challenges, and opportunities in these rapidly evolving fields.
AI-driven companions and therapeutic interfaces in healthcare offer personalized support,
medication reminders, and emotional assistance. These tools, powered by NLP and context-
aware reasoning, improve communication between patients and professionals, potentially
reducing misdiagnoses and enhancing treatment compliance.
AI-driven mental health companions, equipped with NLP algorithms, offer constant,
stigma-free support. While promising in reducing mild to moderate depressive symptoms,
challenges include data privacy, recognizing high-risk scenarios, and avoiding over-
reliance on AI. Despite these concerns, AI companions show potential as scalable mental
health support [13,67].
Socially aware AI systems enhance patient-clinician communication by gathering patient
data, adapting to emotional states, and providing real-time translation. These systems offer
decision-support functions during consultations, augmenting clinician expertise with data-
driven insights. Nevertheless, challenges remain in trust, regulatory compliance, system
interoperability, and cultural sensitivity [11,68].
Education presents another area where ASI's power grows rapidly. AI-driven platforms
support students' intellectual, emotional, and social development. These systems provide
Pag. 272 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

personalized learning, real-time feedback, and inclusivity. Social intelligence in AI tutors

or collaborative learning agents interprets the learners' emotional states and adapts content
delivery. This aims to impart knowledge and develop socio-emotional competencies which
are extremely important in modern life. AI reimagines education as an instructional
assistant and promoter for empathy, cooperation, and broader social skills [69].
Personalized AI tutors in education adapt instruction based on individual student
performance and emotional states, promoting collaborative learning through AI-facilitated
study groups. While offering advantages, this approach raises concerns about over-
personalization, data privacy, and the balance between immediate fulfillment and long-term
cognitive development.
Socially aware AI systems are increasingly used in education to facilitate social skills
training through real-time feedback on interpersonal interactions. These systems that use
techniques like sentiment analysis and facial expression detection offer custom support to
students, particularly to those with special needs, in mastering social cues and peer
interactions. Nevertheless, challenges persist in scalability, contextual nuance, and potential
biases, necessitating ongoing research in order to improve AI systems and develop blended
learning models [58].
Social AI can streamline various operations in corporate environments, from team
communication and project management to customer service and marketing [70]. In
inherently social workplaces, individuals must collaborate, manage hierarchies, and handle
external client interactions. AI agents with social intelligence can facilitate these processes,
boosting productivity, employee satisfaction, and customer engagement. Nevertheless,
businesses deploying such agents must address transparency, privacy, job displacement
concerns, and societal implications of delegating human-facing roles to AI.
Social AI tools, analyzing team communication and interpersonal dynamics, can enhance
project management by identifying miscommunications, optimizing resource allocation,
and improving team morale. However, concerns about surveillance, data consent, and
algorithmic bias must be addressed to ensure ethical implementation [31,71,72].
Socially adept AI, particularly in customer service, can enhance operational efficiency and
brand perception. Advanced virtual agents, equipped with real-time emotion detection and
context-aware decision-making, offer personalized solutions and are able to escalate
complex cases to human representatives. While this hybrid model improves customer
satisfaction and reduces costs, concerns about depersonalization, skill erosion, data
security, and cultural differences persist.
Social AI, particularly recommender systems and digital companions, significantly impacts
daily life through social media and personal interactions. While offering personalized
experiences, it can be both empowering and corrosive, amplifying voices while promoting
echo chambers and enabling manipulative behaviors [35].
Pag. 273 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Recommender systems, while beneficial for personalization, can create echo chambers and
societal polarization by reinforcing existing beliefs. Regulatory efforts and platform
reforms aim to mitigate these effects, but commercial incentives for high engagement pose
a challenge. Social AI extends into personal life through applications that facilitate
relationships and virtual companionship platforms. These platforms use NLP models to
engage users in conversations, offering advice and support, but raise concerns about
authenticity, emotional dependency, and data handling. Despite these challenges, they will
continue to evolve and broaden the perspectives [18,29,61].
As social AI gains traction, its implications for governance and public administration
become more noticeable. Digital platforms and AI-assisted policy making processes change
aspects such as power structures, civic engagement, and accountability. Governments seek
efficiency, transparency, and citizen participation, but these benefits raise fairness, bias,
and democratic norm concerns. Social AI streamlines policy-making and public services,
but it must balance human oversight and community respect [57].
Socially aware citizen engagement platforms, powered by social AI, analyze user
submissions to identify urgent issues and facilitate more inclusive civic dialogues.
Nonetheless, concerns about algorithmic biases, data privacy, and surveillance must be
addressed through careful dataset curation, transparent model design, and robust data
protection protocols [23].
Socially intelligent AI is increasingly used in policymaking, analyzing data to evaluate
public opinion and simulate policy outcomes. While this allows for real-time feedback and
adaptive policymaking, it also raises concerns about policy automation, bias, and lack of
transparency. Addressing these challenges through ethical considerations and democratic
oversight is very important for the responsible use of AI in governance [1,73].

7. Ethical, Legal, and Societal Implications

The rise of ASI has highlighted its ethical, legal, and societal impacts. As AI mimics human
behavior, addressing transparency, privacy, bias, and accountability becomes very
important. While improving technology is important, designing, regulating, and reassessing
the AI ecosystem responsibly is an extremely important necessary step. This section
explores very important questions regarding the integration of AI within modern society. It
emphasizes the need for transparency and explainability in social AI, privacy, data
protection challenges, bias, fairness, inclusivity issues, accountability, legal frameworks,
building trust and ensuring well-being. By understanding these dimensions, AI-driven
transformations can uphold human dignity, autonomy, and progress.
Social AI's urgency is given by its intimacy with individuals and communities. Unlike
abstract computational tasks, social AI agents engage in fluid dialogues, interpret emotions,
and influence relationships. This depth of engagement offers support and enrichment but
Pag. 274 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

also creates risks of manipulation, exploitation, and human value erosion. Ethical dilemmas
intensify when systems replicate empathy or demonstrate emotional responsiveness,
distorting genuine human connection with simulation. Ethical mandates must extend
beyond data management and efficiency to include trust, authenticity, and emotional safety
[11,67,68].
Legal questions arise when AI is embedded in social contexts. AI agents' nuanced
conversations, apparent empathy, and adaptability to social norms suggest they may
participate in social networks rather than being only static tools. This active role presents
novel legal challenges, such as determining if an AI's decision or recommendation carries
the same weight as a human actor's does, establishing liability among developers, data
providers, deployers, and end users, and ensuring alignment with international human rights
standards. Legislators and regulators must balance innovation with potential harms
[5,68,74] .
Societal implications of AI-mediated interaction include transformations in social
structures, cultural norms, and personal identity. Social AI agents can act as companions,
advisors, educators, and intermediaries, potentially redefining family dynamics, reshaping
workplace communication, and changing relationship formation. Some changes may be
positive, encouraging inclusivity and bridging social barriers. Nevertheless, certain social
skills and cultural expressions could erode or be distorted under continuous AI intervention.
Sustained, multidisciplinary research and iterative policy development are needed to
understand these shifts. Addressing the ethical, legal, and societal dimensions of ASI
involves technical design, institutional oversight, and public awareness and engagement
[62].
Transparency and explainability are very important for the responsible use of social AI
systems. While transparency ensures clarity about AI decision-making processes and data
sources, explainability allows AI systems to provide reasoning for their decisions.
Achieving this balance is complex, involving technical solutions, ethical guidelines, and
legal frameworks to ensure accountability and user trust [40].
The adoption of AI agents in social contexts raises privacy concerns due to the collection
and storage of sensitive personal data [57]. While regulatory frameworks like the GDPR
provide legal constraints, the rapid evolution of social AI outpaces legislative processes,
creating ambiguities around compliance. A multi-layered approach involving industry,
government, civil society, and academia is necessary to ensure robust privacy protections,
including transparent communication with users and continuous monitoring of AI systems.
Bias in AI systems, particularly in social AI, arises from unrepresentative or skewed
training data, potentially leading to discrimination and reinforcing harmful social
hierarchies. Addressing these issues requires a complex approach, including data
augmentation, algorithmic fairness, and continuous monitoring. Ensuring inclusivity

Pag. 275 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

involves considering diverse user needs, cultural adaptability, and accessibility, while
strong governance structures and societal conversations are very important for
accountability and informed policymaking [66].
As AI agents become more complex and autonomous, questions of accountability and
liability arise. Determining responsibility for AI-related harm is complex, involving
multiple stakeholders and potentially challenging traditional legal frameworks.
Collaborative efforts among lawmakers, technologists, ethicists, and civil society are
needed to establish robust accountability measures for social AI [66].
Societal acceptance of social AI agents depends on their technical capabilities and users'
trust in their interactions. Trust is linked to AI agents' perceived intentions, reliability,
competence, and morality. As AI integrates into communication, domestic, and
professional settings, people rely on it for support, companionship, decision-making, and
social facilitation. These shifts in behavior and emotions are significant, but concerns about
manipulation, loss of genuine human connection, and psychological dependency arise.
Balancing AI benefits against mental health and social well-being challenges require public
discussions, ethical oversight, and cross-sector collaboration.
Trust is not built solely based on assurances or compliance, it builds through cumulative
interactions and experiences. When an AI assistant consistently offers sound guidance,
accurately interprets emotions, and shows empathy, trust forms. Conversely,
unpredictability, mishandling sensitive information, or expressing biases can erode trust
quickly. In emotionally charged contexts like mental health counseling or grief support, any
breach of privacy or feeling of judgment can cause emotional harm. Once broken, trust is
challenging to rebuild. Rigorous testing, ongoing user feedback, and ethical design
principles are very important [32,40,73].
Psychological well-being and social AI intersect in multiple ways. AI companions and
chatbots can alleviate loneliness, especially for socially isolated individuals. Studies show
potential improvements in mood and perceived social support among older adults using AI-
powered robotic companions [36]. AI systems in education can boost self-esteem with
personalized feedback. Chatbots serve as accessible first-line support in mental health
applications, bridging healthcare gaps. These positive outcomes suggest social AI can
enhance psychological welfare, complementing traditional human interventions [75].
Nevertheless, the pervasive presence of social AI agents may lead to dependencies and
subtle social manipulation. Users might over-rely on AI-generated validation or emotional
support, neglecting human relationships. AI systems gathering psychological profiles might
influence users' behaviors or opinions, raising concerns about autonomy and free will.
Subtle manipulations could be difficult to resist or identify, especially in contexts where
commercial or political interests exploit AI for targeted advertising, information control, or
ideological persuasion. Without clear boundaries and transparent mechanisms, social AI

Pag. 276 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

could inadvertently or deliberately create echo chambers, polarize communities, or

undermine critical thinking [68].
Advancing digital literacy and emotional intelligence among users is very important in
mitigating the risks associated with social AI. Educational initiatives that teach users about
AI's learning, data interpretation, and outcomes empower critical engagement and self-
reflection. Recognizing when AI-driven advice or support should be supplemented by
professionals helps users avoid unhealthy dependencies. Emotional intelligence training for
both users and developers ensures respectful and supportive interactions, minimizing the
risk of harmful manipulation. Robust oversight mechanisms, including independent audits,
ethics review boards, and user advocacy committees, monitor the deployment of AI systems
at scale. Public reporting of key metrics provides insights into how well these systems align
with societal values and psychological needs. Policymakers can mandate reporting, set
transparency requirements, and impose penalties to ensure trust is achieved through
institutional processes and accountability.
Cultural factors shape how communities perceive and integrate social AI. In certain
societies, AI agents facilitating communal problem-solving may be accepted more readily
while more individualist cultures might be cautious about data-sharing and potential
psychological influences. Developers must be sensitive to these cultural dimensions to
design AI that resonates across societies. For instance, language patterns and norms of
polite speech vary widely, and an AI agent's inability to conform can slow down trust and
give rise to misunderstandings [19,50].
Psychological well-being depends on authenticity and genuine connection. Critics argue
that advanced AI companionship can never replicate human emotion and empathy,
potentially diminishing people's motivation and ability to connect with others. Social AI
should serve as a supplement rather than a substitute for human interaction. Proponents
argue that these technologies can fill care and communication gaps when human resources
are insufficient or inaccessible. The debate is ongoing as AI advances to being capable of
nuanced emotional engagement [5,13,68].
A balanced approach is needed to manage social AI's complexities, acknowledging its
potential and risks to trust and psychological health. This includes proactive governance,
user education, rigorous research on long-term impacts, and dialogue between technologists
and the public. Societal trust is built when social AI acts consistently in people's best
interests, respects autonomy, and is governed ethically and legally. Psychological well-
being must remain a priority, ensuring technology contributes to human flourishing rather
than replacing or distorting social connections [19,40].
This exploration of transparency, privacy, bias, accountability, and trust highlights the
depth of ethical, legal, and societal challenges posed by social AI. Integrating AI into human
communication and interaction requires reevaluating norms like data protection and human

Pag. 277 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

agency. Multiple stakeholders must collaborate to control these changes. Responsibility

spans design decisions, policy principles, cultural adaptations, and moral commitments.
Thoughtfully approached, social AI can enrich interaction with empathetic support,
inclusive communication, and complex social understanding. Nevertheless, misuse and
harm highlight the urgency of robust governance and transparent design. In the following,
we explore challenges, research questions, and future opportunities, guiding AI's evolution
towards a more human-centered and socially beneficial paradigm.
8. Challenges and Open Research Questions
The rapid transformation of human interaction by AI agents with varying degrees of ASI
has raised unresolved questions and technical obstacles. Researchers across multiple
disciplines face challenges in conceptualizing, developing, deploying, and governing AI
systems that engage effectively and ethically in complex social environments.

Figure 1. The proposed roadmap highlights a systemic continuity approach rather than an
isolated paired one

Pag. 278 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

This section examines the main challenges and open research questions, including technical
limitations in modeling social phenomena, cultural sensitivity requirements,
interdisciplinary collaboration, security threats from adversarial behavior, and societal
integration. Each dimension poses distinct problems and highlights the inseparability of
technical, social, and ethical considerations in ASI. The proposed roadmap highlights a
systemic continuity approach rather than an isolated paired one, being depicted in Figure 1
above.
Capturing the complexities of real-world social interactions in computational models
remains a persistent challenge for ASI. While progress has been made, many forms of
sociality, such as context-dependent schemas and dynamic group dynamics, remain elusive.
Addressing these challenges requires interdisciplinary collaborations and the development
of advanced computational frameworks that can handle multimodal data, subjectivity, and
uncertainty in real time.
As AI systems become more prevalent, cultural sensitivity becomes very important. This
involves understanding subtle nuances in communication styles, gestures, and social norms
across different cultures. While some aspects of social intelligence may appear universal,
culturally specific adaptations are necessary in order to avoid misinterpretations and ensure
ethical behavior.
ASI requires interdisciplinary collaboration between computer scientists and social
scientists in order to develop meaningful social capabilities. Nevertheless, this collaboration
is hindered by methodological gaps, terminology differences, and institutional structures.
Overcoming these challenges demands proactive strategies, including the creation of multi-
disciplinary teams and the promotion of open dialogue about technical and social scientific
frameworks.
As AI agents become more integrated into human life, security vulnerabilities and
adversarial attacks become more urgent challenges. Adversarial attacks can manipulate data
inputs, exploit trust in AI agents, weaponize generative models, and breach privacy.
Adversarial resilience must be a priority in designing socially intelligent AI systems,
integrating cryptographic safeguards, anomaly detection, robust ML paradigms, and user
education.
Beyond immediate challenges, the long-term trajectory of AI agents in human societies is
profound. As AI matures, it can transform interactions and the broader social, economic,
and political aspects. Nevertheless, the direction and desirability of these transformations
are debated and uncertain. Researchers must address fundamental questions about how AI
will shape norms, structures, and human concepts.
An extremely important area involves the shifting nature of social bonds and interpersonal
relationships. Early evidence suggests users derive emotional support from AI interactions.
These relationships may deepen, raising issues about dependency, authenticity, and
Pag. 279 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

emotional well-being. AI companions could mitigate loneliness, provide empathy, and

encourage personal growth. Nevertheless, they might displace or erode genuine human
connections, especially if AI agents become so compelling that they outcompete human
interaction. The ethical implications of promoting AI-human emotional attachments are
complex, as they can be unilateral, algorithmically mediated, and lacking reciprocity.
Researchers must design AI systems that complement human bonds, supporting socially
beneficial outcomes without undermining communities.
AI's integration into the workplace will have significant consequences. AI systems with
nuanced social interaction may take on roles like recruitment, team management,
negotiation, and customer service. While automation has been discussed in terms of its
economic impact, the social dimension adds complexity. Intelligent agents may perform
emotional labor tasks, such as comforting customers or resolving conflicts, redefining
human skill sets and altering organizational hierarchies. Industries like healthcare,
education, and hospitality may heavily rely on compassionate, empathetic, and culturally
sensitive AI agents. The question is whether these agents will enhance human capabilities
or lead to job displacement and to professional expertise erosion. Policy interventions,
including retraining, labor regulations, and ethical guidelines, will determine how these
technologies coexist with human labor.
The normalization of AI agents in daily life necessitates a deeper exploration of their impact
on social norms and values. Societies evolve through interactions between individuals,
institutions, and traditions. AI integration could accelerate shifts in norms around privacy,
autonomy, and communication. For instance, as people rely on AI assistants for social
interactions, expectations of responsiveness, politeness, and emotional availability may
change. This could lead to new etiquette frameworks for the interactions between human
beings and AI agents. Alternatively, it could exacerbate social inequalities as those with
advanced AI tools better manage complex social networks. The trajectory depends on
various factors, including technological innovation, market forces, cultural receptivity,
legal frameworks, and public preferences.
AI's role in shaping collective decision-making processes is a significant open question.
Some envision AI as a facilitator of "deliberative democracy", providing data-driven
insights and moderating debates. This could reduce polarization by highlighting areas of
agreement. Nevertheless, critics argue that ceding control of civic positions to algorithmic
mediators risks undermining democratic values, especially if models are controlled by
vested interests. Biases in training data or manipulations in information presentation can
steer public opinion and amplify divisions. Rigorous research is needed to evaluate how AI
systems might inadvertently institutionalize biases or marginalize certain voices.
Developing transparent, accountable frameworks for deploying socially intelligent agents
in public decision-making is very important.

Pag. 280 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Long-term integration hinges on the AI's ability to adapt ethically over time. Human
societies continually redefine ethical boundaries. AI agents must have mechanisms for
ethical self-reflection, guided by human oversight or normative frameworks. This raises the
possibility of designing "moral machines" with dynamic ethical reasoning. Nevertheless,
philosophical and technical hurdles remain. Determining ethical standards and resolving
conflicts among competing moral systems are unresolved questions with global
implications. Ensuring transparency and democratic oversight is necessary for maintaining
trust.
Ecological sustainability and social stratification are very important aspects of evaluating
the long-term role of socially intelligent AI. Training and deploying large-scale AI systems
consumes significant computational resources and non-renewable energy, potentially
exacerbating environmental costs and colliding with international emissions reduction
commitments. If AI integration into daily life increases, energy demand could soar.
Researchers must design energy-efficient models, explore decentralized computing, and
integrate sustainability into AI ethics. Failing to do so risks accelerating the climate crisis.
Technological adoption also leads to social stratification, with wealthier communities
benefiting sooner. If socially intelligent AI determines social and professional success,
existing inequalities may worsen. Targeted interventions like subsidized or open-source AI
solutions can mitigate disparities, but their effectiveness is uncertain. An integrative
unifying perspective considering race, gender, age, and socioeconomic status is needed for
equitable long-term strategies.
The future trajectory of AI in human societies requires sustained, globally inclusive
discussions. Research should extend beyond laboratories to longitudinal field studies, pilot
projects, and continuous feedback from diverse user groups. Cross-cultural collaboration
involving global tech companies, local innovators, and community leaders is very important
in order to ensure that AI integration does not exacerbate historical injustices or undermine
cultural autonomy. The interaction of technology, policy, business interests, and civic
values will determine whether AI evolves into a tool of liberation or social control.
The proposed roadmap comprising the identified challenges in ASI along with key insights
towards solutions proposals is synthetized in Table 1 below.

Challenge The Reason for its Key Insights Towards

Importance Solutions Proposals

Modeling Complex Current approaches Multi-modal sensors, richer

Social Phenomena struggle with sequential models and
high-dimensional, continual-learning pipelines
sequential and
culture-laden behavior
Pag. 281 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Cultural and Context Risk of mishandling Transfer/meta-learning plus

Sensitivity idioms, norms or genuinely participatory
under-represented groups data-collection to "learn with"
local communities

Interdisciplinary and Social science insight Shared vocabularies, hybrid

Methodological Gaps rarely reaches model qualitative-quantitative
design or evaluation benchmarks, new collaboratory
metrics funding models

Security and Adversarial Social bots are attractive Secure multi-agent protocols,
Threats targets, adversarial inputs game-theoretic defenses and
can hijack narratives "safety-first" reward shaping

Societal and Ethical Fragmented regulation, Global policy road-maps,

Governance opacity erodes trust and transparent auditing,
accountability user-centric consent flows

Sustainability and Carbon cost of large Energy-efficient architectures,

Equity models and the "digital open-source subsidies,
divide" threaten inclusive intersectional impact audits
progress

Human and Artificial Long-term vision where Embodied co-learning systems,

Intelligence Symbiosis AI augments rather than mixed-reality collaboration,
supplants human agency continual ethical co-adaptation

Table 1. Proposed Roadmap Comprising the Identified Challenges in ASI along with Key
Insights Towards Solutions Proposals
The long-term integration of AI agents in human societies presents a defining challenge,
encompassing technical, ethical, economic, and existential considerations. The ultimate
shape and impact of these systems remain uncertain, and research plans must adapt to AI's
connection with human needs and aspirations. Societies' evolution towards benevolent AI
partners enhancing human flourishing or dystopian scenarios involving pervasive
surveillance and social fragmentation depends on current decisions. The task's complexity
emphasizes the urgency of robust, ethically grounded research, transparent policymaking,
and inclusive societal debates.
The challenges and open research questions surrounding ASI and its impact on human
interaction are vast and complex. Technical hurdles in modeling social phenomena, the
need for cultural sensitivity, interdisciplinary collaboration, adversarial attacks, and long-
Pag. 282 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

term integration implications demand sustained attention from various stakeholders. Only
comprehensive, ethically informed, and collaborative efforts can harness the potential of
socially intelligent AI for the collective well-being while respecting human social
complexities.

9. Future Directions and Opportunities

ASI, a dynamic research area, bases on knowledge from diverse disciplines like computer
science, cognitive psychology, and ethics. As AI integrates into human social spaces, its
impact expands. Controlling the design, development, and deployment of AI with technical
complexity and sensitivity to human emotions and cultures is very important. Ongoing
research, policy-making, and cross-disciplinary engagement ensure responsible evolution
and unlock novel possibilities. Emerging technologies for enhanced social cognition,
ethical frameworks, and policy roadmaps, generalized social intelligence, human-AI
symbiosis, along with a vision for the coming decades and beyond are explored.
The development of ASI relies on advanced technologies that enhance AI agents' social
perception, interpretation, and responsiveness. These technologies include complex
sensors, mixed reality interfaces, edge computing, quantum computing, neuromorphic
computing, and brain-computer interfaces. By leveraging these technologies, ASI aims to
create AI agents capable of near-human levels of social interaction and understanding in
complex environments.
Ethical considerations and regulatory initiatives are very important as AI agents gain social
responsibilities and decision-making capacities. Explainability, data privacy, bias, and
fairness are dimensions that demand rigorous analysis. Policymaking must be adaptive,
incorporating multi-stakeholder collaboration and global discussions in order to maximize
the potential of social AI while safeguarding public welfare and individual rights.
Achieving generalized social intelligence in AI involves creating agents capable of adapting
to diverse tasks and contexts. This requires multi-modal and multi-cultural competence,
integrating language understanding, cultural anthropology, and advanced ML architectures.
The approach involves interdisciplinary efforts, flexible architectures, and ethical
considerations in order to develop AI agents that can enhance human communication and
cross-cultural collaboration.
As AI integrates into society, a key question arises about genuine symbiosis between
humans and artificial entities. Symbiosis goes beyond cooperation, it is a relationship where
humans and AI enhance each other's capabilities and well-being, leading to new forms of
collective intelligence and creativity. Advances in personalization, adaptive learning, trust-
building, along with merging physical and digital realities are important prerequisites.

Pag. 283 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Personalized interaction is very important for symbiosis. AI systems that accurately model
individual preferences and cognitive patterns can tailor assistance and companionship. In
education, symbiotic AI tutors collaborate with teachers and students, providing just-in-
time support that augments skills or emotional resilience. Over time, these adaptive tutors
co-create educational content, refining lessons based on student feedback, while teachers
maintain oversight. This collaborative learning system combines human intuition and
empathy with the AI's data-driven insights.
Trust is very important for symbiosis, but humans may be hesitant to trust AI partners due
to concerns about data misuse, opaque decision-making, or hidden programs. Socially
intelligent AI designed for symbiotic relationships should prioritize transparency,
consistency, and reliability. These systems should acknowledge their limitations, clarify
uncertainty, and provide users with additional resources or expert opinions. Continuous
user feedback loops reinforce trust by granting humans agency in shaping the AI's role.
"Explainable symbiosis" may emerge, where AI provides correct or contextually relevant
outputs to deepen mutual understanding by specifying core reasoning or data patterns.
Symbiosis implies growth and transformation for both humans and AI. For instance, AI-
driven brainstorming assistants in creative industries can propose diverse concepts,
enabling artists and designers to explore new aesthetic territories. Over time, this symbiotic
relationship can shape human perception of creative processes, leading to emergent forms
of expression. Similarly, AI can serve as a partner for hypothesis generation, data synthesis,
or strategic planning in knowledge work, freeing human collaborators to focus on higher-
level decision-making or ethical considerations. In these scenarios, AI becomes a valued
co-contributor whose inputs and insights are integrated into the human mental model.
Physical embodiments of AI, like social robots, exoskeletons, or prosthetics, augment
human physical capabilities. Socially intelligent exoskeletons could aid rehabilitation by
dynamically adjusting support levels based on progress and factors like terrain or fatigue.
These robotic aids, integrated with the user's body schema, could become extensions of
their sense of agency. Over time, human intention and AI-driven movement can lead to
augmented embodiment, diminishing the biological and technological boundaries.
Organizational structures and social norms will also change. AI-enabled group decision-
making platforms collect inputs, identify consensus, and suggest compromise solutions.
Socially aware AI in boardrooms, strategic planning, or public policy forums can make use
of collaborative intelligence by integrating multiple perspectives. These AI systems manage
conflict, recognize agreement or dissent, and adapt communication styles to de-escalate
tension or build rapport. This leads to more inclusive and transparent processes that
empower marginalized voices. Human–AI symbiosis amplifies collective intelligence, but
it also raises questions about responsibility and leadership when AI insights are the ones
shaping the outcomes.

Pag. 284 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Achieving true symbiosis between humans and AI faces challenges. Over-reliance on AI

may diminish essential skills and critical thinking. AI agents with cultural or social biases
could reinforce those biases, creating a symbiotic loop. Power imbalances in AI access may
lead to stratified symbiosis, benefiting some while others lag. Addressing these pitfalls
requires thoughtful design, oversight, resource distribution, and public debates on AI
boundaries.
Despite complexities, the symbiosis between humans and AI technology remains a
powerful motivator for social AI research. Focusing on personalization, trust, co-creation,
augmented embodiment, and organizational transformation can lead to AI as a genuine
partner, not just a tool or competitor. Mutual respect, aligning AI with human values, and
refining collaborative processes are key to success. Symbiosis is not an unattainable ideal
but a practical, human-machine evolving partnership.
The future of social AI, driven by advancements in machine learning and the interaction
between human and computer technology, promises to integrate seamlessly into daily life,
healthcare, education, and beyond. While this integration offers opportunities for
personalized care, adaptive learning, and enhanced community engagement, it also raises
concerns about privacy, bias, and significant changes in social interactions. Ultimately, the
responsible development and deployment of social AI will be very important in creating a
future where technology amplifies human creativity, empathy, and problem-solving
capacity.

10. Conclusions
ASI, rapidly impacting human interaction, is a complex domain where computational
systems replicate, simulate, and reconfigure social dynamics. Unlike just processing
information or solving logical tasks, true ASI arises from integrating psychological,
cognitive, and sociocultural insights into ML architectures, NLP, computer vision, multi-
agent systems, and reinforcement learning. These advancements enable AI agents, from
chatbots to advanced social robots, to interpret, predict, and shape human behavior in real-
time.
The conducted analysis highlighted the interaction between AI theoretical constructs and
practical implementations. Traditional AI relied on rigid rule-based systems or specialized
ML models for limited interaction. Recent advances in deep learning, transfer learning, and
multi-modal processing expanded the AI's operational domain, enabling new forms of
interaction. Conversational agents understand contextual shifts, computer vision systems
detect facial expressions, and reinforcement learning agents adapt in dynamic settings.
Examining these technological aspects shows that AI is evolving from technical problem-
solvers to social participants, reinforcing the need for social intelligence combining
reasoning, perception, adaptability, and cultural sensitivity.
Pag. 285 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

AI's deployment in social contexts spans decades, starting with primitive chatbots like
ELIZA, progressing to interactive robots, and culminating in contemporary virtual
assistants managing personalized services. These historical aspects highlight the tension
between aspirations and realizations, where early visions faced limitations in sustaining
natural dialogues or interpreting emotions. Recent improvements in GPU computing, large-
scale NNs, and datasets have addressed these limitations. Today, AI-driven systems are
integrated into daily life, from digital customer service to healthcare companionship. These
developments change our perception of AI agents from static tools to dynamic interactants
in the human social existence.
Throughout the article, transformations in communication and group dynamics were
highlighted. AI-mediated interactions reconfigure social exchange patterns, altering
conversational flow and agency distribution. AI agents can facilitate group interactions as
neutral mediators, but they can also introduce biases, reinforce echo chambers, or
undermine privacy. These shifts in social norms and etiquette necessitate recalibrating
social protocols in AI-coexistent environments. Boundaries between private conversations
and AI-stored data tend to dissipate therefore challenging autonomy, disclosure, and
intimacy assumptions. The role of AI in empathy adoption or hindrance is a concern, as
mechanized empathy can extend care or lead to manipulation.
Socially aware AI has proven beneficial in diverse domains, from education and healthcare
to corporate and governmental settings. In education, personalized tutoring systems
improve engagement, while group-based AI facilitates collaborative learning. In healthcare,
social robots and AI companions support mental health, covering the needs of isolated
populations. Corporates use AI for project management and team communication,
sometimes replacing traditional management tasks. Governments and public administration
employ AI to engage citizens, reflecting an effort to integrate intelligent platforms into
governance. Nevertheless, accountability, data security, and the potential erosion of direct
human oversight remain pertinent concerns.
AI's ethical, legal, and societal implications are very important aspects, especially when it
influences mental health, political stance, or group decision-making. Bias in AI-driven
interactions arises from skewed training data or flawed algorithms, emphasizing the
importance of inclusive design processes. Privacy and data protection concerns grow as
social AI collects sensitive data. Accountability mechanisms, from legal frameworks to
public oversight, are emerging but lag behind innovation. Societal acceptance depends on
ethical behavior, privacy protection, fairness, and security against manipulation.
AI is evolving into a social actor that supports, augments, or supplants human interaction.
Interdisciplinary integration between computer science, psychology, sociology, ethics, and
law is very important for constructing general models that are capturing human
communication and group behavior. While remarkable milestones have been reached,

Pag. 286 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

fundamental challenges remain, including modeling context, culture, and emotion, and
societal questions about autonomy, identity, and authenticity in AI-augmented spaces.
The article emphasizes the need for critical engagement with AI systems that assume social
roles. AI's evolving complexity holds significant potential to change the way in which
individuals and societies construct meaning, form relationships, and organize collective
activities. The pursuit of artificial social intelligence is no longer a peripheral research
question but a central concern requiring robust theoretical models, empirical validation,
ethical scrutiny, and inclusive design practices. An integrative approach is needed to ensure
that AI's growing presence in human social life serves as a technical milestone, a channel
for human flourishing, and an opportunity to enrich, rather than diminish, the collective
social experience.
ASI's implications extend to various stakeholders, shaping how individuals, communities,
institutions, and global networks interact with AI-augmented realities. In the case of users,
socially aware AI offers personalization, convenience, and accessibility, addressing
emotional states, interpersonal nuances, and contextual cues. Mental health support is
accessible through empathetic AI companions, while adaptive tutoring and therapy cater to
social and cognitive impairments. In consumer contexts, AI assistants enhance daily tasks
with nuanced social understanding, from scheduling to domestic management.
Nevertheless, AI's implications for individuals are mixed. While convenience is appealing,
AI systems trained on proprietary data continuously gather insights into personal habits,
emotions, and social networks, potentially eroding privacy and autonomy. Surveillance
demonstrates how data can be harvested and leveraged for targeted advertising, potentially
harming users' interests. Psychologically, relying on AI for social needs raises questions
about how human empathy and emotional intelligence might decline if AI mediates more
interpersonal contact. AI's subtle influence on self-expression, self-awareness, and social
development must be carefully considered in order to avoid offloading human capacities to
machines.
At the community and societal levels, socially adept AI offers opportunities for collective
problem-solving, inclusive group dynamics, and enhanced public discussions. Social robots
in eldercare facilitate group activities and reduce isolation, while AI-based platforms in
education provide broader and more equitable access to specialized knowledge and
collaborative learning experiences. AI-driven interfaces can simplify bureaucratic
processes, promote digital town halls, deliver multilingual support, increase civic
participation and lower barriers to engagement. These improvements can be significant in
under-resourced or remote settings where access to human experts or specialized
infrastructure is scarce.
Society-wide implications reveal complex aspects. AI-mediated social systems can
systematically manipulate public sentiment through content curation, targeted persuasion,

Pag. 287 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

or synthetic media, making institutions susceptible to algorithmic influence. Socially

oriented AI can either promote civic renewal or exacerbate political polarization. When AI
becomes a cultural gatekeeper, limiting public discourse to automated curation, modelling
social norms, values, and discussions become very important aspects to analyze. AI's
proliferation in social media, news aggregation, and interpersonal communication
emphasizes this significance.
In the case of corporate stakeholders, AI-driven social environments and processes promise
efficiency, insights, and innovative service delivery. Customer service deploys
conversational agents for routine inquiries and data collection. Decisional factors analyze
internal communication for collaboration bottlenecks and team dynamics improvements.
Cost savings, faster decision-making, and a cohesive culture entice integration.
Nevertheless, ethical dilemmas arise from employee surveillance, data ownership, and job
displacement. The modern workplace's change to AI-driven decision-making raises labor
questions about error responsibility, inclusive workforce development, and preserving
human qualities like trust and empathy in professional contexts.
Policymakers face unique challenges in responding to AI's profound social transformations.
Drafting and implementing regulations that balance innovation, and public welfare requires
forward-looking legal frameworks for rapid technological changes. Social intelligence in
AI diminishes the user and subject separation, raising questions of liability, data
governance, and ethical accountability. Explainability is very important in public
administration, where citizens must trust resource distribution, service administration, and
policy recommendations. Transparency, oversight, and cross-sectoral partnerships maintain
democratic accountability when AI reshapes public opinion, participation, and resource
allocation. Governments risk widening social disparities, enabling discrimination, or
authoritarian surveillance.
Collaborative, interdisciplinary engagement is important for socially aware AI systems with
ethical principles, robust design, and cultural sensitivity. Societal transformations will vary
based on resources, infrastructure, norms, and regulations, magnifying global inequities.
Stakeholders responsible for fairness, explainability, and inclusivity can harness the AI's
positive potential while mitigating adverse outcomes. Open dialogue between
technologists, ethicists, user communities, corporations, and public institutions is very
important, in situations where each sector contributes expertise to shape AI attuned to
human values and social well-being.
Despite significant progress in developing AI systems with social intelligence, intrinsic
limitations persist due to technical, methodological, and conceptual challenges. Capturing
the contextual richness of human social interaction is particularly difficult. Human
communication involves various signals, and while AI excels at language parsing, facial
expression recognition, and context-awareness, it struggles to adapt to rapidly shifting
environments with subtle contextual cues, power dynamics, historical relationships, and
Pag. 288 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

cultural norms. The high variability of social scenarios means that impressive capabilities
in a well-defined domain can be inadequate in novel or cross-cultural interactions. Models
trained in one language or cultural context often fail to generalize, highlighting the
challenge of building truly universal social intelligence.
Modern AI's data-driven foundations face limitations. Deep learning systems require large,
high-quality training datasets representing real-world interactions. Nevertheless, these
datasets often suffer from biases, inconsistencies, or insufficient coverage, leading to
harmful stereotypes and misinterpretations of underrepresented groups. Mitigating bias
through specialized algorithms or representative datasets is partial and addressing how
societal inequities are encoded in data and replicated by AI is challenging. Quantitative
measures like accuracy do not capture qualitative dimensions of social intelligence like
trust-building, empathy, or moral discernment.
Methodologically, there is a significant need for integrating theoretical models from
psychology, sociology, anthropology, and related fields into AI architectures. While some
research attempts to import concepts like theory of mind, self-awareness, or moral
reasoning into algorithmic frameworks, these are nascent and often fragmented.
Reinforcement learning, for instance, trains agents to optimize rewards for cooperative
behavior but fails to grasp ethical dimensions or long-term social implications. Multi-agent
systems that cooperate or compete in dynamic environments are hindered by modeling
emergent social phenomena like complex group behaviors, hierarchical roles, and intricate
relationship networks. Achieving a rigorous, interdisciplinary methodology that unites
social science insights with computational efficiency remains an open research frontier.
Technical limitations in explainability, interpretability, and real-time adaptation affect
black-box models, especially large NNs. These models can produce socially compelling
responses but lack transparency in their decision-making processes. This opacity is
problematic in high-stakes contexts like healthcare diagnosis or criminal justice risk
assessments, undermining user trust. While explainable AI research advances, reconciling
interpretability with complex neural architectures still remains challenging. Real-time
adaptation requires continuous context updates, potentially leading to re-training or
catastrophic forgetting. Achieving adaptability and stability is very important, especially in
multi-party interactions where trust, timing, and rapport building are of extreme
importance.
The ethical and regulatory frameworks remain fragmented. Existing guidelines, like
responsible AI, highlight broad principles but lack granularity to address socially capable
AI deployment. Issues of consent, data ownership, and user autonomy are challenging when
AI integrates into daily life, capturing sensitive social data. Without cohesive legal
frameworks, the risk of misuse can become large. Normative debates on AI integration in
domains like childcare, therapy, eldercare, and education remain unresolved. Preserving

Pag. 289 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

human agency and ensuring AI advances human dignity rather than infringing on it is both
technical and philosophical.
Adversarial behavior and security vulnerabilities pose challenges for AI systems operating
in open social contexts. Malicious actors can manipulate outputs, disseminate
misinformation, or exploit system weaknesses. Adversarial examples, specially crafted
inputs that deceive AI models, demonstrate that advanced models remain susceptible to
tampering. Targeting social AI systems with adversarial methods can spread disinformation
and undermine public trust. Safeguarding AI from these threats requires specialized
countermeasures, but few comprehensive frameworks exist to robustly defend socially
oriented AI applications.
These limitations highlight the challenges of achieving genuine ASI. While progress has
been made, the envisioned environment where AI seamlessly manages human social life is
still far from realized. This aspect emphasizes the need for patient, thorough research that
integrates engineering ingenuity with humanistic insight to design AI that aligns with social
reality. Only through sustained effort can the field move closer to producing AI systems
that truly enrich human life, rather than merely replicating social interaction.
Artificial social intelligence holds immense potential. AI agents' integration into human
society through language comprehension, emotional attune, adaptive behavior, and
collaborative intelligence suggests a new era of human-machine interactions. These
developments promise enhanced efficiency, improved access to services, amplified
creativity, and societal benefits in education, healthcare, governance, and personal well-
being. Nevertheless, they also raise concerns about privacy erosion, social divisions,
manipulation, and displacement of empathy, trust, and mutual understanding.
Technologists, policymakers, ethicists, psychologists, and diverse voices from
marginalized communities, industry leaders, government agencies, and academic
researchers must collaborate to develop socially responsible AI. This collaboration should
establish shared principles, including clear regulations protecting individuals' rights,
incentives for ethical AI design, and public engagement campaigns to clarify the AI's
societal impact. Participatory processes ensure a genuine alignment with global societies'
pluralistic values.
Incorporating interdisciplinary research methodologies from the humanities and social
sciences into technological design is another very important aspect. Engineering solutions
to social interaction challenges should be tested against psychological theories of emotion,
cognition, and interpersonal dynamics. Anthropology can guide AI system development by
understanding cultural variation in communication styles and normative behaviors.
Sociology and political science can reveal macro-level implications of AI deployment, such
as power changes, identity formation, and civic participation. By integrating social science

Pag. 290 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

theories, AI developers can move beyond superficial human interaction emulations and
create systems that engage with and replicate human social complexity.
Initiatives to adopt transparency, explainability, and accountability must continue and
expand. While explainable AI research makes progress, further transparency around data
usage, decision-making, and system limitations is very important for building user trust and
responsible AI deployment. Governments and international bodies must collaborate to
create enforceable standards for comprehensible AI system reporting. Accountability is
urgent as AI systems operate semi-autonomously in sensitive domains. Deciding where
responsibility lies, with developers, data sources, deployers, or users, is a complex problem.
Legal frameworks must evolve to address responsibility distribution when AI becomes a
participant in human social ecosystems.
Robust ethical oversight and governance are strategic for sustaining societal trust. Ethical
review boards or cross-disciplinary committees can integrate into AI research pipelines to
systematically identify and mitigate potential risks. Oversight mechanisms must adapt to
AI technology evolution, reflecting real-time updates in capabilities, data collection, and
usage. Civic organizations and public play a very important role in scrutinizing AI
deployments, supporting user rights, and ensuring equitable AI adoption.
Improving public literacy in AI is very important as social interactions with AI become
commonplace. Educational programs can empower citizens to make informed decisions
about data sharing, chatbot engagement, and cautious use of AI. Encouraging digital
skepticism helps inoculate society against AI manipulation. Improvements in user interface
design and consent processes make interactions more transparent, preserving individual
agency.
A broader vision of the symbiosis between humans and AI technology emerges, envisioning
collaborative, mutually beneficial relationships that enhance human capacities while
preserving autonomy, dignity, and sociability. This symbiosis requires ongoing negotiation
involving technical breakthroughs, policy changes, and cultural discussions about human
social bonds. Ensuring that AI augments human potential rather than subjugating it is
fundamental. The AI's role as a social actor should complement human emotional
intelligence, ethical reasoning, and creative expression, recognizing the depth of human
lived experience while contributing computational insights to solve the society's problems.
Without a doubt, ASI requires a multi-layered approach. Researchers and developers should
integrate cross-disciplinary wisdom, refine computational models for cultural and ethical
sensitivity, and strive for new forms of transparency and user engagement. Policymakers
and regulatory agencies should collaborate internationally to establish frameworks that
safeguard public interests and encourage responsible innovation. Corporate stakeholders
should view ethical, social, and cultural considerations as strategic requirements, investing
in comprehensive risk assessments and community engagement. Civil society groups,

Pag. 291 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

educators, and media organizations should raise public awareness about the AI's impact on
society. Finally, individual users should remain vigilant, recognizing that every interaction
with AI shapes the norms and values of technologically mediated social life. The future
trajectory of AI depends on collective efforts to design, deploy, regulate, and use AI
ethically and responsibly. By promoting collaboration, ethical rigor, and human sociality,
we can steer AI toward outcomes that enhance societal flourishing and promote human
identity. The path forward requires vigilance, imagination, and a commitment to shaping
technology for humanity's best interests.

Acknowledgment
The authors would like to express their gratitude for the logistics support received from the
Center of Research, Consultancy and Training in Economic Informatics and Information
Technology RAU-INFORTIS of the Romanian-American University and the Center for
Computational Science and Machine Intelligence of the Romanian-American University.

References
[1] Lim, T. Environmental, Social, and Governance (ESG) and Artificial Intelligence in
Finance: State-of-the-Art and Research Takeaways. Artif Intell Rev 2024, 57,
doi:10.1007/s10462-024-10708-3.
[2] Sun, T.Q. Adopting Artificial Intelligence in Public Healthcare: The Effect of Social
Power and Learning Algorithms. Int J Environ Res Public Health 2021, 18,
doi:10.3390/ijerph182312682.
[3] St. Amant, R.; Young, R.M. Links: Artificial Intelligence and Interactive
Entertainment. Intelligence 2001, 12, doi:10.1145/378116.378120.
[4] Bharadiya, J. Artificial Intelligence in Transportation Systems A Critical Review.
American Journal of Computing and Engineering 2023, 6, doi:10.47672/ajce.1487.
[5] Brännström, A.; Wester, J.; Nieves, J.C. A Formal Understanding of Computational
Empathy in Interactive Agents. Cogn Syst Res 2024, 85,
doi:10.1016/j.cogsys.2023.101203.
[6] Uludag, K. The Use of AI-Supported Chatbot in Psychology. SSRN Electronic Journal
2023, doi:10.2139/ssrn.4331367.
[7] Baranowski, M. The Database Construction of Reality in the Age of AI: The Coming
Revolution in Sociology? AI Soc 2024, doi:10.1007/s00146-024-01873-8.
Pag. 292 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[8] Kartal, G.; Yeşilyurt, Y.E. A Bibliometric Analysis of Artificial Intelligence in L2

Teaching and Applied Linguistics between 1995 and 2022. ReCALL 2024,
doi:10.1017/S0958344024000077.
[9] Montag, C.; Ali, R.; Davis, K.L. Affective Neuroscience Theory and Attitudes towards
Artificial Intelligence. AI Soc 2024, doi:10.1007/s00146-023-01841-8.
[10] Van Rooij, I.; Guest, O.; Adolfi, F.; De Haan, R.; Kolokolova, A.; Rich, P.
Reclaiming AI as a Theoretical Tool for Cognitive Science. PsyArXiv 2023.
[11] Ciraolo, D.; Fazio, M.; Calabrò, R.S.; Villari, M.; Celesti, A. Facial Expression
Recognition Based on Emotional Artificial Intelligence for Tele-Rehabilitation. Biomed
Signal Process Control 2024, 92, doi:10.1016/j.bspc.2024.106096.
[12] Miranda, L.; Garibary, O.O. Approaching (Super)Human Intent Recognition in Stag
Hunt with the Naïve Utility Calculus Generative Model. Comput Math Organ Theory
2023, 29, doi:10.1007/s10588-022-09367-y.
[13] Sharma, A.; Lin, I.W.; Miner, A.S.; Atkins, D.C.; Althoff, T. Human–AI
Collaboration Enables More Empathic Conversations in Text-Based Peer-to-Peer Mental
Health Support. Nat Mach Intell 2023, 5, doi:10.1038/s42256-022-00593-2.
[14] Chang, Y.L.; Ke, J. Socially Responsible Artificial Intelligence Empowered People
Analytics: A Novel Framework Towards Sustainability. Human Resource Development
Review 2024, 23, doi:10.1177/15344843231200930.
[15] Li, J.; Chu, Y.; Xu, J. Impression Transference from AI to Human: The Impact of
AI’s Fairness on Interpersonal Perception in AI-Mediated Communication. International
Journal of Human Computer Studies 2023, 179, doi:10.1016/j.ijhcs.2023.103119.
[16] Celik, A.; Eltawil, A.M. At the Dawn of Generative AI Era: A Tutorial-Cum-Survey
on New Frontiers in 6G Wireless Intelligence. IEEE Open Journal of the Communications
Society 2024, 5, doi:10.1109/OJCOMS.2024.3362271.
[17] Panicker, A.; Sasidevan, V. Social Adaptive Behavior and Oscillatory Prevalence in
an Epidemic Model on Evolving Random Geometric Graphs. Chaos Solitons Fractals
2024, 178, doi:10.1016/j.chaos.2023.114407.
[18] Gloor, P.; Fronzetti Colladon, A.; Grippa, F. Measuring Ethical Behavior with AI and
Natural Language Processing to Assess Business Success. Sci Rep 2022, 12,
doi:10.1038/s41598-022-14101-4.
[19] Tiwari, R. Ethical And Societal Implications of AI and Machine Learning.
INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND
MANAGEMENT 2023, 07, doi:10.55041/ijsrem17519.

Pag. 293 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[20] Prajapati, D.K.; Upadhayay, D.; Kumawat, A.; Shekhawat, A.P. Application of AI
Computer Vision and Image Recognition. Industrial Engineering Journal 2022, 51,
doi:10.36893/iej.2022.v51i10.081-087.
[21] Baabdullah, A.M. Generative Conversational AI Agent for Managerial Practices: The
Role of IQ Dimensions, Novelty Seeking and Ethical Concerns. Technol Forecast Soc
Change 2024, 198, doi:10.1016/j.techfore.2023.122951.
[22] Broadbent, E.; Loveys, K.; Ilan, G.; Chen, G.; Chilukuri, M.M.; Boardman, S.G.;
Doraiswamy, P.M.; Skuler, D. ElliQ, an AI-Driven Social Robot to Alleviate Loneliness:
Progress and Lessons Learned. The Journal of Aging Research & Lifestyle 2024,
doi:10.14283/jarlife.2024.2.
[23] González, A.S.; Rampino, L. A Design Perspective on How to Tackle Gender Biases
When Developing AI-Driven Systems. AI and Ethics 2024, doi:10.1007/s43681-023-
00386-2.
[24] Singh, A.; Kanaujia, A.; Singh, V.K.; Vinuesa, R. Artificial Intelligence for
Sustainable Development Goals: Bibliometric Patterns and Concept Evolution
Trajectories. Sustainable Development 2024, 32, doi:10.1002/sd.2706.
[25] Li, J.; Wu, L.; Qi, J.; Zhang, Y.; Wu, Z.; Hu, S. Determinants Affecting Consumer
Trust in Communication With AI Chatbots: The Moderating Effect of Privacy Concerns.
Journal of Organizational and End User Computing 2023, 35,
doi:10.4018/JOEUC.328089.
[26] Farhan, A. The Impact of Artificial Intelligence on Human Workers. Journal Of
Communication Education 2023, 17, doi:10.58217/joce-ip.v17i2.350.
[27] Coman, A.W.; Cardon, P. Perceptions of Professionalism and Authenticity in AI-
Assisted Writing. Business and Professional Communication Quarterly 2024,
doi:10.1177/23294906241233224.
[28] Liu, J.; Li, C.; Pan, J.; Guo, J. Visual Communication of Moving Images Based on AI
Recognition and Light Sensing Image Edge Detection Algorithm. Opt Quantum Electron
2024, 56, doi:10.1007/s11082-024-06542-0.
[29] Ismail, W.S. Threat Detection and Response Using AI and NLP in Cybersecurity.
Journal of Internet Services and Information Security 2024, 14,
doi:10.58346/JISIS.2024.I1.013.
[30] Neupane, S.; Mitra, S.; Fernandez, I.A.; Saha, S.; Mittal, S.; Chen, J.; Pillai, N.;
Rahimi, S. Security Considerations in AI-Robotics: A Survey of Current Methods,
Challenges, and Opportunities. IEEE Access 2024, 12,
doi:10.1109/ACCESS.2024.3363657.
[31] Alexander Obaigbena; Oluwaseun Augustine Lottu; Ejike David Ugwuanyi; Boma
Sonimitiem Jacks; Enoch Oluwademilade Sodiya; Obinna Donald Daraojimba;
Pag. 294 / 350
Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Oluwaseun Augustine Lottu AI and Human-Robot Interaction: A Review of Recent

Advances and Challenges. GSC Advanced Research and Reviews 2024, 18,
doi:10.30574/gscarr.2024.18.2.0070.
[32] Braunschweig, B.; Buijsman, S.; Chamroukhi, F.; Heintz, F.; Khomh, F.; Mattioli, J.;
Poretschkin, M. AITA: AI Trustworthiness Assessment. AI and Ethics 2024, 4,
doi:10.1007/s43681-023-00397-z.
[33] Booyse, D.; Scheepers, C.B. Barriers to Adopting Automated Organisational
Decision-Making through the Use of Artificial Intelligence. Management Research
Review 2024, 47, doi:10.1108/MRR-09-2021-0701.
[34] Norzelan, N.A.; Mohamed, I.S.; Mohamad, M. Technology Acceptance of Artificial
Intelligence (AI) among Heads of Finance and Accounting Units in the Shared Service
Industry. Technol Forecast Soc Change 2024, 198, doi:10.1016/j.techfore.2023.123022.
[35] Wei, Y.; Tyson, G. Understanding the Impact of AI Generated Content on Social
Media: The Pixiv Case. 2024, doi:10.1145/3664647.3680631.
[36] Oladoyinbo, T.O.; Olabanji, S.O.; Olaniyi, O.O.; Adebiyi, O.O.; Okunleye, O.J.;
Alao, A.I. Exploring the Challenges of Artificial Intelligence in Data Integrity and Its
Influence on Social Dynamics. Asian Journal of Advanced Research and Reports 2024,
18, doi:10.9734/ajarr/2024/v18i2601.
[37] Voruganti, S.S.J.; Gadepalli, T.N. Person-Hood Issues Related To Artificial
Intelligence. International Journal For Multidisciplinary Research 2024, 6,
doi:10.36948/ijfmr.2024.v06i01.14177.
[38] Sufyan, N.S.; Fadhel, F.H.; Alkhathami, S.S.; Mukhadi, J.Y.A. Artificial Intelligence
and Social Intelligence: Preliminary Comparison Study between AI Models and
Psychologists. Front Psychol 2024, 15, doi:10.3389/fpsyg.2024.1353022.
[39] Otis, N.; Clarke, R.; Delecourt, S.; Holtz, D.; Koning, R. The Uneven Impact of
Generative AI on Entrepreneurial Performance. SSRN Electronic Journal 2024,
doi:10.2139/ssrn.4671369.
[40] Oloyede, J. Ethical Reflections on AI for Cybersecurity: Building Trust. SSRN
Electronic Journal 2024, doi:10.2139/ssrn.4733563.
[41] Rafique, H.; Nazeer, I.; Rehman, J. The Impact of ChatGPT on Language Evolution:
A Linguistic Analysis. Journal of Education and Social Studies 2024, 5,
doi:10.52223/jess.2024.5106.
[42] Li, J.; Huang, J.; Liu, J.; Zheng, T. Human-AI Cooperation: Modes and Their Effects
on Attitudes. Telematics and Informatics 2022, 73, doi:10.1016/j.tele.2022.101862.

Pag. 295 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[43] Abdul Wasay Khan Mandokhail The Transformative Role of Artificial Intelligence in
Conflict Resolution and Peacekeeping. NUST Journal of International Peace & Stability
2024, doi:10.37540/njips.v7i1.166.
[44] Bird, J.J.; Lotfi, A. CIFAKE: Image Classification and Explainable Identification of
AI-Generated Synthetic Images. IEEE Access 2024, 12,
doi:10.1109/ACCESS.2024.3356122.
[45] Peng, Y.; Han, J.; Zhang, Z.; Fan, L.; Liu, T.; Qi, S.; Feng, X.; Ma, Y.; Wang, Y.; Zhu,
S.C. The Tong Test: Evaluating Artificial General Intelligence Through Dynamic
Embodied Physical and Social Interactions. Engineering 2024, 34,
doi:10.1016/j.eng.2023.07.006.
[46] Babu, A.; Boddu, S.B. BERT-Based Medical Chatbot: Enhancing Healthcare
Communication through Natural Language Understanding. Exploratory Research in
Clinical and Social Pharmacy 2024, 13, doi:10.1016/j.rcsop.2024.100419.
[47] Kalogiannidis, S.; Kalfas, D.; Papaevangelou, O.; Giannarakis, G.;
Chatzitheodoridis, F. The Role of Artificial Intelligence Technology in Predictive Risk
Assessment for Business Continuity: A Case Study of Greece. Risks 2024, 12,
doi:10.3390/risks12020019.
[48] Park, S.; Lee, J. yoon; Lee, J. AI System Architecture Design Methodology Based on
IMO (Input-AI Model-Output) Structure for Successful AI Adoption in Organizations.
Data Knowl Eng 2024, 150, doi:10.1016/j.datak.2023.102264.
[49] Sai, S.; Garg, A.; Jhawar, K.; Chamola, V.; Sikdar, B. A Comprehensive Survey on
Artificial Intelligence for Unmanned Aerial Vehicles. IEEE Open Journal of Vehicular
Technology 2023, 4, doi:10.1109/OJVT.2023.3316181.
[50] Simon, J.; Rieder, G.; Branford, J. The Philosophy and Ethics of AI: Conceptual,
Empirical, and Technological Investigations into Values. Digital Society 2024, 3,
doi:10.1007/s44206-024-00094-2.
[51] Chongwatpol, J. A Technological, Data-Driven Design Journey for Artificial
Intelligence (AI) Initiatives. Educ Inf Technol (Dordr) 2024, doi:10.1007/s10639-024-
12459-8.
[52] Lenaerts, T.; Saponara, M.; Pacheco, J.M.; Santos, F.C. Evolution of a Theory of
Mind. iScience 2024, 27, doi:10.1016/j.isci.2024.108862.
[53] Harré, M.S.; El-Tarifi, H. Testing Game Theory of Mind Models for Artificial
Intelligence. Games (Basel) 2024, 15, doi:10.3390/g15010001.
[54] Cheng, C.; Wang, C.; Yang, D.; Wen, X.; Liu, W.; Zhang, F. Underwater Small Target
Detection Based on Dynamic Convolution and Attention Mechanism. Front Mar Sci
2024, 11, doi:10.3389/fmars.2024.1348883.

Pag. 296 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[55] Lahudeo, B.N.; Vayadande, M.; Malviya, R.; Haldule, A. Walking and Survival AI
Using Reinforcement Learning - Simulation. International Journal of Scientific Research
in Computer Science, Engineering and Information Technology 2024, 10,
doi:10.32628/cseit2390629.
[56] Pirnau, M.; Botezatu, M.A.; Priescu, I.; Hosszu, A.; Tabusca, A.; Coculescu, C.;
Oncioiu, I. Content Analysis Using Specific Natural Language Processing Methods for
Big Data. Electronics (Switzerland) 2024, 13, doi:10.3390/electronics13030584.
[57] Kaleel, A.; Alomari, M.S. Integrating Artificial Intelligence in Public Relations and
Media: A Bibliometric Analysis of Emerging Trends and Influences. Iraqi Journal for
Computer Science and Mathematics 2024, 5, doi:10.52866/ijcsm.2024.05.01.002.
[58] Krugmann, J.O.; Hartmann, J. Sentiment Analysis in the Age of Generative AI.
Customer Needs and Solutions 2024, 11, doi:10.1007/s40547-024-00143-4.
[59] Shaikh Mohd Azhar Mohd Abrar Integrating Artificial Intelligence with Human
Psychology. International Journal of Advanced Research in Science, Communication and
Technology 2024, doi:10.48175/ijarsct-15250.
[60] Morais da Silva, D.; Da Costa Farias, R.; Cunha, A.; Salete Casagrande, L.; Antunes
Peres, R. History and Legacy of Alan Turing for Computer Science. International Journal
of Scientific Research and Management (IJSRM) 2024, 12,
doi:10.18535/ijsrm/v12i02.ec06.
[61] Zhou, B.; Yang, G.; Shi, Z.; Ma, S. Natural Language Processing for Smart
Healthcare. IEEE Rev Biomed Eng 2024, 17, doi:10.1109/RBME.2022.3210270.
[62] Vanhée, L. Organizing Transdisciplinary Research and Innovation: The Case of the
Socially-Aware Artificial Intelligence Focus Area. In Proceedings of the CEUR Workshop
Proceedings; 2024; Vol. 3634.
[63] Farrelly, T.; Baker, N. Generative Artificial Intelligence: Implications and
Considerations for Higher Education Practice. Educ Sci (Basel) 2023, 13.
[64] Chandra, B.; Rahman, Z. Artificial Intelligence and Value Co-Creation: A Review,
Conceptual Framework and Directions for Future Research. Journal of Service Theory
and Practice 2024, 34, doi:10.1108/JSTP-03-2023-0097.
[65] Zhou, M.; Lei, L.; Chen, W.; Luo, Q.; Li, J.; Zhou, F.; Yang, X.; Pan, Y. Deep
Learning-Based Diagnosis of Aortic Dissection Using an Electrocardiogram:
Development, Validation, and Clinical Implications of the AADE Score. Kardiol Pol
2024, 82, doi:10.33963/v.phj.98880.
[66] Flathmann, C.; McNeese, N.J.; Schelble, B.; Knijnenburg, B.; Freeman, G.
Understanding the Impact and Design of AI Teammate Etiquette. Hum Comput Interact
2024, 39, doi:10.1080/07370024.2023.2189595.

Pag. 297 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[67] Elyoseph, Z.; Refoua, E.; Asraf, K.; Lvovsky, M.; Shimoni, Y.; Hadar-Shoval, D.
Capacity of Generative AI to Interpret Human Emotions From Visual and Textual Data:
Pilot Evaluation Study. JMIR Ment Health 2024, 11, doi:10.2196/54369.
[68] Liu, W.; Zhang, S.; Zhang, T.; Gu, Q.; Han, W.; Zhu, Y. The AI Empathy Effect: A
Mechanism of Emotional Contagion. Journal of Hospitality Marketing and Management
2024, 33, doi:10.1080/19368623.2024.2315954.
[69] Kim, J.; Ham, Y.; Lee, S.S. Differences in Student-AI Interaction Process on a
Drawing Task: Focusing on Students’ Attitude towards AI and the Level of Drawing
Skills. Australasian Journal of Educational Technology 2024, 40, doi:10.14742/ajet.8859.
[70] Zhang, R.; Duan, W.; Flathmann, C.; Mcneese, N.; Freeman, G.; Williams, A.
Investigating AI Teammate Communication Strategies and Their Impact in Human-AI
Teams for Effective Teamwork. Proc ACM Hum Comput Interact 2023, 7,
doi:10.1145/3610072.
[71] Hu, Y.H.; Fu, J.S.; Yeh, H.C. Developing an Early-Warning System through Robotic
Process Automation: Are Intelligent Tutoring Robots as Effective as Human Teachers?
Interactive Learning Environments 2023, doi:10.1080/10494820.2022.2160467.
[72] Deng, W.; Liu, Q.; Zhao, F.; Pham, D.T.; Hu, J.; Wang, Y.; Zhou, Z. Learning by
Doing: A Dual-Loop Implementation Architecture of Deep Active Learning and Human-
Machine Collaboration for Smart Robot Vision. Robot Comput Integr Manuf 2024, 86,
doi:10.1016/j.rcim.2023.102673.
[73] Laux, J.; Wachter, S.; Mittelstadt, B. Trustworthy Artificial Intelligence and the
European Union AI Act: On the Conflation of Trustworthiness and Acceptability of Risk.
Regul Gov 2024, 18, doi:10.1111/rego.12512.
[74] Kleinrichert, D. Empathy: An Ethical Consideration of AI & Others in the
Workplace. AI Soc 2024, doi:10.1007/s00146-023-01831-w.
[75] Kusal, S.D.; Patil, S.G.; Choudrie, J.; Kotecha, K. V. Understanding the Performance
of AI Algorithms in Text-Based Emotion Detection for Conversational Agents. ACM
Transactions on Asian and Low-Resource Language Information Processing 2024, 23,
doi:10.1145/3643133.

Bibliography
Abdul Wasay Khan Mandokhail The Transformative Role of Artificial Intelligence in
Conflict Resolution and Peacekeeping. NUST Journal of International Peace & Stability
2024, doi:10.37540/njips.v7i1.166.

Pag. 298 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Alexander Obaigbena; Oluwaseun Augustine Lottu; Ejike David Ugwuanyi; Boma

Sonimitiem Jacks; Enoch Oluwademilade Sodiya; Obinna Donald Daraojimba;
Oluwaseun Augustine Lottu AI and Human-Robot Interaction: A Review of Recent
Advances and Challenges. GSC Advanced Research and Reviews 2024, 18,
doi:10.30574/gscarr.2024.18.2.0070.

Baabdullah, A.M. Generative Conversational AI Agent for Managerial Practices: The

Role of IQ Dimensions, Novelty Seeking and Ethical Concerns. Technol Forecast Soc
Change 2024, 198, doi:10.1016/j.techfore.2023.122951.

Babu, A.; Boddu, S.B. BERT-Based Medical Chatbot: Enhancing Healthcare

Communication through Natural Language Understanding. Exploratory Research in
Clinical and Social Pharmacy 2024, 13, doi:10.1016/j.rcsop.2024.100419.

Baranowski, M. The Database Construction of Reality in the Age of AI: The Coming
Revolution in Sociology? AI Soc 2024, doi:10.1007/s00146-024-01873-8.

Bharadiya, J. Artificial Intelligence in Transportation Systems A Critical Review.

American Journal of Computing and Engineering 2023, 6, doi:10.47672/ajce.1487.

Bird, J.J.; Lotfi, A. CIFAKE: Image Classification and Explainable Identification of AI-
Generated Synthetic Images. IEEE Access 2024, 12,
doi:10.1109/ACCESS.2024.3356122.

Booyse, D.; Scheepers, C.B. Barriers to Adopting Automated Organisational Decision-

Making through the Use of Artificial Intelligence. Management Research Review 2024,
47, doi:10.1108/MRR-09-2021-0701.

Brännström, A.; Wester, J.; Nieves, J.C. A Formal Understanding of Computational

Empathy in Interactive Agents. Cogn Syst Res 2024, 85,
doi:10.1016/j.cogsys.2023.101203.

Braunschweig, B.; Buijsman, S.; Chamroukhi, F.; Heintz, F.; Khomh, F.; Mattioli, J.;
Poretschkin, M. AITA: AI Trustworthiness Assessment. AI and Ethics 2024, 4,
doi:10.1007/s43681-023-00397-z.

Broadbent, E.; Loveys, K.; Ilan, G.; Chen, G.; Chilukuri, M.M.; Boardman, S.G.;
Doraiswamy, P.M.; Skuler, D. ElliQ, an AI-Driven Social Robot to Alleviate Loneliness:
Progress and Lessons Learned. The Journal of Aging Research & Lifestyle 2024,
doi:10.14283/jarlife.2024.2.

Celik, A.; Eltawil, A.M. At the Dawn of Generative AI Era: A Tutorial-Cum-Survey on

New Frontiers in 6G Wireless Intelligence. IEEE Open Journal of the Communications
Society 2024, 5, doi:10.1109/OJCOMS.2024.3362271.

Pag. 299 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Chandra, B.; Rahman, Z. Artificial Intelligence and Value Co-Creation: A Review,

Conceptual Framework and Directions for Future Research. Journal of Service Theory
and Practice 2024, 34, doi:10.1108/JSTP-03-2023-0097.

Chang, Y.L.; Ke, J. Socially Responsible Artificial Intelligence Empowered People

Analytics: A Novel Framework Towards Sustainability. Human Resource Development
Review 2024, 23, doi:10.1177/15344843231200930.

Chaturvedi, R.; Verma, S.; Das, R.; Dwivedi, Y.K. Social Companionship with Artificial
Intelligence: Recent Trends and Future Avenues. Technol Forecast Soc Change 2023, 193,
doi:10.1016/j.techfore.2023.122634.

Cheng, C.; Wang, C.; Yang, D.; Wen, X.; Liu, W.; Zhang, F. Underwater Small Target
Detection Based on Dynamic Convolution and Attention Mechanism. Front Mar Sci
2024, 11, doi:10.3389/fmars.2024.1348883.

Chongwatpol, J. A Technological, Data-Driven Design Journey for Artificial Intelligence

(AI) Initiatives. Educ Inf Technol (Dordr) 2024, doi:10.1007/s10639-024-12459-8.

Ciraolo, D.; Fazio, M.; Calabrò, R.S.; Villari, M.; Celesti, A. Facial Expression
Recognition Based on Emotional Artificial Intelligence for Tele-Rehabilitation. Biomed
Signal Process Control 2024, 92, doi:10.1016/j.bspc.2024.106096.

Coman, A.W.; Cardon, P. Perceptions of Professionalism and Authenticity in AI-Assisted

Writing. Business and Professional Communication Quarterly 2024,
doi:10.1177/23294906241233224.

Deng, W.; Liu, Q.; Zhao, F.; Pham, D.T.; Hu, J.; Wang, Y.; Zhou, Z. Learning by Doing: A
Dual-Loop Implementation Architecture of Deep Active Learning and Human-Machine
Collaboration for Smart Robot Vision. Robot Comput Integr Manuf 2024, 86,
doi:10.1016/j.rcim.2023.102673.

Elyoseph, Z.; Refoua, E.; Asraf, K.; Lvovsky, M.; Shimoni, Y.; Hadar-Shoval, D. Capacity
of Generative AI to Interpret Human Emotions From Visual and Textual Data: Pilot
Evaluation Study. JMIR Ment Health 2024, 11, doi:10.2196/54369.

Farhan, A. The Impact of Artificial Intelligence on Human Workers. Journal Of

Communication Education 2023, 17, doi:10.58217/joce-ip.v17i2.350.

Farrelly, T.; Baker, N. Generative Artificial Intelligence: Implications and Considerations

for Higher Education Practice. Educ Sci (Basel) 2023, 13.

Flathmann, C.; McNeese, N.J.; Schelble, B.; Knijnenburg, B.; Freeman, G. Understanding
the Impact and Design of AI Teammate Etiquette. Hum Comput Interact 2024, 39,
doi:10.1080/07370024.2023.2189595.

Pag. 300 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Gloor, P.; Fronzetti Colladon, A.; Grippa, F. Measuring Ethical Behavior with AI and
Natural Language Processing to Assess Business Success. Sci Rep 2022, 12,
doi:10.1038/s41598-022-14101-4.

González, A.S.; Rampino, L. A Design Perspective on How to Tackle Gender Biases

When Developing AI-Driven Systems. AI and Ethics 2024, doi:10.1007/s43681-023-
00386-2.

Harré, M.S.; El-Tarifi, H. Testing Game Theory of Mind Models for Artificial
Intelligence. Games (Basel) 2024, 15, doi:10.3390/g15010001.

Hu, Y.H.; Fu, J.S.; Yeh, H.C. Developing an Early-Warning System through Robotic
Process Automation: Are Intelligent Tutoring Robots as Effective as Human Teachers?
Interactive Learning Environments 2023, doi:10.1080/10494820.2022.2160467.

Huang, L.; Freeman, J.; Cooke, N.J.; Cohen, M.C.; Yin, X.; Clark, J.; Wood, M.;
Buchanan, V.; Corral, C.; Scholcover, F.; et al. Establishing Human Observer Criterion in
Evaluating Artificial Social Intelligence Agents in a Search and Rescue Task. Top Cogn
Sci 2023, doi:10.1111/tops.12648.

Ismail, W.S. Threat Detection and Response Using AI and NLP in Cybersecurity. Journal
of Internet Services and Information Security 2024, 14, doi:10.58346/JISIS.2024.I1.013.

Johnston, W.J.; Cortez, R.M. Business-to-Business Digitalization, Artificial Intelligence,

and Social Action. J Bus Res 2024, 172.

Kaleel, A.; Alomari, M.S. Integrating Artificial Intelligence in Public Relations and
Media: A Bibliometric Analysis of Emerging Trends and Influences. Iraqi Journal for
Computer Science and Mathematics 2024, 5, doi:10.52866/ijcsm.2024.05.01.002.

Kalogiannidis, S.; Kalfas, D.; Papaevangelou, O.; Giannarakis, G.; Chatzitheodoridis, F.

The Role of Artificial Intelligence Technology in Predictive Risk Assessment for Business
Continuity: A Case Study of Greece. Risks 2024, 12, doi:10.3390/risks12020019.

Kartal, G.; Yeşilyurt, Y.E. A Bibliometric Analysis of Artificial Intelligence in L2

Teaching and Applied Linguistics between 1995 and 2022. ReCALL 2024,
doi:10.1017/S0958344024000077.

Kim, J.; Ham, Y.; Lee, S.S. Differences in Student-AI Interaction Process on a Drawing
Task: Focusing on Students’ Attitude towards AI and the Level of Drawing Skills.
Australasian Journal of Educational Technology 2024, 40, doi:10.14742/ajet.8859.

Kleinrichert, D. Empathy: An Ethical Consideration of AI & Others in the Workplace. AI

Soc 2024, doi:10.1007/s00146-023-01831-w.

Pag. 301 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Krugmann, J.O.; Hartmann, J. Sentiment Analysis in the Age of Generative AI. Customer
Needs and Solutions 2024, 11, doi:10.1007/s40547-024-00143-4.

Kusal, S.D.; Patil, S.G.; Choudrie, J.; Kotecha, K. V. Understanding the Performance of
AI Algorithms in Text-Based Emotion Detection for Conversational Agents. ACM
Transactions on Asian and Low-Resource Language Information Processing 2024, 23,
doi:10.1145/3643133.

Lahudeo, B.N.; Vayadande, M.; Malviya, R.; Haldule, A. Walking and Survival AI Using
Reinforcement Learning - Simulation. International Journal of Scientific Research in
Computer Science, Engineering and Information Technology 2024, 10,
doi:10.32628/cseit2390629.

Laux, J.; Wachter, S.; Mittelstadt, B. Trustworthy Artificial Intelligence and the European
Union AI Act: On the Conflation of Trustworthiness and Acceptability of Risk. Regul Gov
2024, 18, doi:10.1111/rego.12512.

Lenaerts, T.; Saponara, M.; Pacheco, J.M.; Santos, F.C. Evolution of a Theory of Mind.
iScience 2024, 27, doi:10.1016/j.isci.2024.108862.

Li, J.; Chu, Y.; Xu, J. Impression Transference from AI to Human: The Impact of AI’s
Fairness on Interpersonal Perception in AI-Mediated Communication. International
Journal of Human Computer Studies 2023, 179, doi:10.1016/j.ijhcs.2023.103119.

Li, J.; Huang, J.; Liu, J.; Zheng, T. Human-AI Cooperation: Modes and Their Effects on
Attitudes. Telematics and Informatics 2022, 73, doi:10.1016/j.tele.2022.101862.

Li, J.; Wu, L.; Qi, J.; Zhang, Y.; Wu, Z.; Hu, S. Determinants Affecting Consumer Trust in
Communication With AI Chatbots: The Moderating Effect of Privacy Concerns. Journal
of Organizational and End User Computing 2023, 35, doi:10.4018/JOEUC.328089.

Lim, T. Environmental, Social, and Governance (ESG) and Artificial Intelligence in

Finance: State-of-the-Art and Research Takeaways. Artif Intell Rev 2024, 57,
doi:10.1007/s10462-024-10708-3.

Liu, J.; Li, C.; Pan, J.; Guo, J. Visual Communication of Moving Images Based on AI
Recognition and Light Sensing Image Edge Detection Algorithm. Opt Quantum Electron
2024, 56, doi:10.1007/s11082-024-06542-0.

Liu, W.; Zhang, S.; Zhang, T.; Gu, Q.; Han, W.; Zhu, Y. The AI Empathy Effect: A
Mechanism of Emotional Contagion. Journal of Hospitality Marketing and Management
2024, 33, doi:10.1080/19368623.2024.2315954.

Pag. 302 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Miranda, L.; Garibary, O.O. Approaching (Super)Human Intent Recognition in Stag Hunt
with the Naïve Utility Calculus Generative Model. Comput Math Organ Theory 2023, 29,
doi:10.1007/s10588-022-09367-y.

Montag, C.; Ali, R.; Davis, K.L. Affective Neuroscience Theory and Attitudes towards
Artificial Intelligence. AI Soc 2024, doi:10.1007/s00146-023-01841-8.

Morais da Silva, D.; Da Costa Farias, R.; Cunha, A.; Salete Casagrande, L.; Antunes
Peres, R. History and Legacy of Alan Turing for Computer Science. International Journal
of Scientific Research and Management (IJSRM) 2024, 12,
doi:10.18535/ijsrm/v12i02.ec06.

Neupane, S.; Mitra, S.; Fernandez, I.A.; Saha, S.; Mittal, S.; Chen, J.; Pillai, N.; Rahimi,
S. Security Considerations in AI-Robotics: A Survey of Current Methods, Challenges, and
Opportunities. IEEE Access 2024, 12, doi:10.1109/ACCESS.2024.3363657.

Norzelan, N.A.; Mohamed, I.S.; Mohamad, M. Technology Acceptance of Artificial

Intelligence (AI) among Heads of Finance and Accounting Units in the Shared Service
Industry. Technol Forecast Soc Change 2024, 198, doi:10.1016/j.techfore.2023.123022.

Oladoyinbo, T.O.; Olabanji, S.O.; Olaniyi, O.O.; Adebiyi, O.O.; Okunleye, O.J.; Alao,
A.I. Exploring the Challenges of Artificial Intelligence in Data Integrity and Its Influence
on Social Dynamics. Asian Journal of Advanced Research and Reports 2024, 18,
doi:10.9734/ajarr/2024/v18i2601.

Oloyede, J. Ethical Reflections on AI for Cybersecurity: Building Trust. SSRN Electronic

Journal 2024, doi:10.2139/ssrn.4733563.

Otis, N.; Clarke, R.; Delecourt, S.; Holtz, D.; Koning, R. The Uneven Impact of
Generative AI on Entrepreneurial Performance. SSRN Electronic Journal 2024,
doi:10.2139/ssrn.4671369.

Panicker, A.; Sasidevan, V. Social Adaptive Behavior and Oscillatory Prevalence in an

Epidemic Model on Evolving Random Geometric Graphs. Chaos Solitons Fractals 2024,
178, doi:10.1016/j.chaos.2023.114407.

Park, S.; Lee, J. yoon; Lee, J. AI System Architecture Design Methodology Based on
IMO (Input-AI Model-Output) Structure for Successful AI Adoption in Organizations.
Data Knowl Eng 2024, 150, doi:10.1016/j.datak.2023.102264.

Peng, Y.; Han, J.; Zhang, Z.; Fan, L.; Liu, T.; Qi, S.; Feng, X.; Ma, Y.; Wang, Y.; Zhu, S.C.
The Tong Test: Evaluating Artificial General Intelligence Through Dynamic Embodied
Physical and Social Interactions. Engineering 2024, 34, doi:10.1016/j.eng.2023.07.006.

Pag. 303 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Pirnau, M.; Botezatu, M.A.; Priescu, I.; Hosszu, A.; Tabusca, A.; Coculescu, C.; Oncioiu,
I. Content Analysis Using Specific Natural Language Processing Methods for Big Data.
Electronics (Switzerland) 2024, 13, doi:10.3390/electronics13030584.

Prajapati, D.K.; Upadhayay, D.; Kumawat, A.; Shekhawat, A.P. Application of AI

Computer Vision and Image Recognition. Industrial Engineering Journal 2022, 51,
doi:10.36893/iej.2022.v51i10.081-087.

Rafique, H.; Nazeer, I.; Rehman, J. The Impact of ChatGPT on Language Evolution: A
Linguistic Analysis. Journal of Education and Social Studies 2024, 5,
doi:10.52223/jess.2024.5106.

Sai, S.; Garg, A.; Jhawar, K.; Chamola, V.; Sikdar, B. A Comprehensive Survey on
Artificial Intelligence for Unmanned Aerial Vehicles. IEEE Open Journal of Vehicular
Technology 2023, 4, doi:10.1109/OJVT.2023.3316181.

Shaikh Mohd Azhar Mohd Abrar Integrating Artificial Intelligence with Human
Psychology. International Journal of Advanced Research in Science, Communication and
Technology 2024, doi:10.48175/ijarsct-15250.

Sharma, A.; Lin, I.W.; Miner, A.S.; Atkins, D.C.; Althoff, T. Human–AI Collaboration
Enables More Empathic Conversations in Text-Based Peer-to-Peer Mental Health
Support. Nat Mach Intell 2023, 5, doi:10.1038/s42256-022-00593-2.

Simon, J.; Rieder, G.; Branford, J. The Philosophy and Ethics of AI: Conceptual,
Empirical, and Technological Investigations into Values. Digital Society 2024, 3,
doi:10.1007/s44206-024-00094-2.

Singh, A.; Kanaujia, A.; Singh, V.K.; Vinuesa, R. Artificial Intelligence for Sustainable
Development Goals: Bibliometric Patterns and Concept Evolution Trajectories.
Sustainable Development 2024, 32, doi:10.1002/sd.2706.

St. Amant, R.; Young, R.M. Links: Artificial Intelligence and Interactive Entertainment.
Intelligence 2001, 12, doi:10.1145/378116.378120.

Sufyan, N.S.; Fadhel, F.H.; Alkhathami, S.S.; Mukhadi, J.Y.A. Artificial Intelligence and
Social Intelligence: Preliminary Comparison Study between AI Models and
Psychologists. Front Psychol 2024, 15, doi:10.3389/fpsyg.2024.1353022.

Sun, T.Q. Adopting Artificial Intelligence in Public Healthcare: The Effect of Social
Power and Learning Algorithms. Int J Environ Res Public Health 2021, 18,
doi:10.3390/ijerph182312682.

Pag. 304 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Tiwari, R. Ethical And Societal Implications of AI and Machine Learning.

INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND
MANAGEMENT 2023, 07, doi:10.55041/ijsrem17519.

Uludag, K. The Use of AI-Supported Chatbot in Psychology. SSRN Electronic Journal

2023, doi:10.2139/ssrn.4331367.

Van Rooij, I.; Guest, O.; Adolfi, F.; De Haan, R.; Kolokolova, A.; Rich, P. Reclaiming AI
as a Theoretical Tool for Cognitive Science. PsyArXiv 2023.

Vanhée, L. Organizing Transdisciplinary Research and Innovation: The Case of the

Socially-Aware Artificial Intelligence Focus Area. In Proceedings of the CEUR Workshop
Proceedings; 2024; Vol. 3634.

Voruganti, S.S.J.; Gadepalli, T.N. Person-Hood Issues Related To Artificial Intelligence.

International Journal For Multidisciplinary Research 2024, 6,
doi:10.36948/ijfmr.2024.v06i01.14177.

Wei, Y.; Tyson, G. Understanding the Impact of AI Generated Content on Social Media:
The Pixiv Case. 2024, doi:10.1145/3664647.3680631.

Wided, R.; Alfalih, A.A. Extremism Immunity through Artificial Intelligence Networks:
Extremism Awareness and Social Intelligence. International Journal of Data and Network
Science 2023, 7, doi:10.5267/j.ijdns.2022.9.013.

Yetişensoy, O.; Rapoport, A. Artificial Intelligence Literacy Teaching in Social Studies

Education. Journal of Pedagogical Research 2023, 7, doi:10.33902/JPR.202320866.

Zhang, R.; Duan, W.; Flathmann, C.; Mcneese, N.; Freeman, G.; Williams, A.
Investigating AI Teammate Communication Strategies and Their Impact in Human-AI
Teams for Effective Teamwork. Proc ACM Hum Comput Interact 2023, 7,
doi:10.1145/3610072.

Zhou, B.; Yang, G.; Shi, Z.; Ma, S. Natural Language Processing for Smart Healthcare.
IEEE Rev Biomed Eng 2024, 17, doi:10.1109/RBME.2022.3210270.

Zhou, M.; Lei, L.; Chen, W.; Luo, Q.; Li, J.; Zhou, F.; Yang, X.; Pan, Y. Deep Learning-
Based Diagnosis of Aortic Dissection Using an Electrocardiogram: Development,
Validation, and Clinical Implications of the AADE Score. Kardiol Pol 2024, 82,
doi:10.33963/v.phj.98880.

Pag. 305 / 350

Article’s total number of pages: 47
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

GENERATING JAVA CODE WITH AI TOOLS. USAGE AND IMPLICATIONS

Alexandru TĂBUȘCĂ1
Andrei LUCHICI2
Mihai BOTEZATU3
Silvia TĂBUȘCĂ4

Abstract
Java programming, a high-level, general-purpose language renowned for its "Write Once,
Run Anywhere" (WORA) capability, has (re)gained notable traction as developers
increasingly integrate artificial intelligence (AI) generative tools into their workflows.
Java’s platform independence, robust security features, and extensive libraries have made
it a preferred choice for a wide range of applications, from mobile apps to large-scale
enterprise systems. The advent of AI generative tools, such as ChatGPT, GitHub Copilot or
Amazon CodeWhisperer, has further enhanced Java programming by automating mundane
tasks, improving code quality, and fostering creativity in the development process. In
today’s world, solid knowledge related to AI code generation tools is a must for all
developers and software engineers. AI tools for generating Java code have also started an
entire new set of debates related to copyright issues. Currently, the relevant legal
frameworks, at international level, are not harmonized and in some cases even antagonistic.
Keywords: ai code generation; java ai tools; java; ai code legal issues; ai copyright issues
JEL Classification: C8, O31, O33, O34, O39

1. Introduction
Java programming, a high-level, general-purpose language renowned for its "Write Once,
Run Anywhere" (WORA) capability, has gained notable traction as developers increasingly
integrate artificial intelligence (AI) generative tools into their workflows. Java’s platform
independence, its’ very robust security features, as well as the extensive libraries have made
it an excellent and preferred choice for a very wide range of different applications, starting
with mobile apps and going to large-scale enterprise systems. The advent of AI generative
tools, such as GitHub Copilot and Amazon CodeWhisperer, has further enhanced Java

1PhD, Associate Professor, Romanian-American University, Romania, alex.tabusca@rau.ro, corresponding

author
2PhD, Lecturer, Center for Research in AI, Romania, andrei.luchici@rau.ro
3PhD, Professor Habil, Romanian-American University, Romania, mihai.botezatu@rau.ro
4PhD, Lecturer, Center for Human Rights and Migration, Romania, silvia.tabusca@rau.ro

Pag. 306 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

programming by automating mundane tasks, improving (usually) the quality of code, and
increasing creativity within the entire software development process [1][2][3][4]. The
integration of AI generative tools is reshaping the software development landscape,
promoting efficiency and enabling developers to focus on complex problem-solving. These
tools assist in various capacities, including code generation, debugging, and project
documentation, thereby streamlining workflows and enhancing productivity [5][6][7].
Moreover, studies have shown a positive correlation between the use of these AI tools and
perceived productivity among developers, as they consolidate multiple functions into a
single application and provide quick access to critical information [5][8]. However, the
incorporation of AI into Java programming is not without its challenges. Developers face
issues related to the reliability of generated outputs, the need for precise prompting, and
concerns regarding data security and governance. Ethical considerations, such as
algorithmic bias and intellectual property rights, further complicate the landscape as the
technology matures [9][10][11]. As the Java community continues to embrace AI solutions,
it is imperative to address these challenges through transparency, ethical guidelines, and
best practices to ensure that the advantages of AI can be fully realized without
compromising quality or security [12][13][14]. Looking to the future, the collaboration
between AI and developers is expected to deepen, with increasing use cases for AI in Java
development, particularly in specialized areas such as natural language processing and
dynamic pricing systems. As developers adapt to these evolving technologies, ongoing
education and a commitment to ethical standards will be crucial in harnessing the full
potential of AI while mitigating associated risks [15][16][14].

2. Java Programming Paradigm

Java is a high-level, general-purpose programming language that is both memory-safe and
object-oriented, making it a prominent choice among software developers. Designed with
the principle of "Write Once, Run Anywhere" (WORA), Java allows compiled code to run
on any platform that supports Java without needing recompilation. This is achieved through
the use of the Java Virtual Machine (JVM), which interprets Java bytecode across different
operating systems and hardware architectures [1][2].

2.1. Features of Java

Object-Oriented Principles
Java is fundamentally an object-oriented programming (OOP) language, supporting core
OOP concepts such as classes, objects, inheritance, encapsulation, and polymorphism.
Unlike some other programming languages, all code in Java is encapsulated within classes,
and every data item is treated as an object, except for the primitive data types which are
optimized for performance [3][2]. The syntax of Java is influenced heavily by C and C++,
Pag. 307 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

yet it avoids complex features like pointers and multiple inheritance, promoting simplicity
and enhancing its usability for beginners [4][17].
Platform Independence
One of Java's most significant advantages is its platform independence. Java programs are
first compiled into bytecode, a platform-neutral intermediate representation. This bytecode
can run on any system equipped with a JVM, making Java applications remarkably versatile
across different environments [3][2][17].
Performance and Scalability
Despite being an interpreted language, Java's performance is impressive, thanks in part to
its Just-In-Time (JIT) compiler, which optimizes bytecode execution. Additionally, Java
supports multithreading, allowing concurrent execution of code, which improves efficiency
and resource management [17][18]. These features, combined with Java's inherent
scalability, make it suitable for a wide range of applications from mobile devices to
enterprise-level systems [4][17].
Security and Robustness
Java is designed with security in mind, incorporating features that protect against common
programming errors, such as memory leaks and buffer overflows. Its strong type-checking
at compile time and runtime contributes to its reliability and robustness, ensuring that Java
applications are both secure and stable [3][2]. The Java ecosystem includes a rich set of
libraries for various tasks, including internationalization, database connectivity, and remote
method invocation, further enhancing its utility for developers [1][4].
Development Environment
The Java development environment is extensive, comprising several frameworks and
libraries. The Java Standard Edition (Java SE) provides the core functionality, while Java
Enterprise Edition (Java EE) extends this with capabilities for building large-scale,
distributed applications [1]. Various tools and integrated development environments (IDEs)
facilitate coding in Java, allowing developers to leverage the language's features effectively
in their projects [30].

3. AI Generative Tools for Java Programming

AI generative tools have become instrumental in enhancing the efficiency and effectiveness
of Java programming. These tools leverage artificial intelligence to assist developers in
various capacities, streamlining workflows and improving code quality.
The adoption (more or less generalized at this time) of AI in code generation offers several
benefits, including faster development cycles, improved productivity, and the ability to
automate mundane tasks. Nevertheless, challenges persist, particularly concerning the
Pag. 308 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

quality of the generated code and potential security risks. Developers must remain vigilant
against issues such as security vulnerabilities that may arise from AI-generated outputs,
which can lead to exploitable weaknesses in software systems [45][31]. Moreover, the
copyright implications of using AI-generated code are under scrutiny. As AI tools produce
outputs based on their training datasets—which may include copyrighted material - the
legal status of the code generated raises questions about ownership and intellectual property
rights [46][43]. The intersection of AI-generated content and copyright law continues to
evolve, necessitating further examination of how these technologies interact with existing
legal frameworks. The advent of artificial intelligence (AI) tools for code generation has
prompted significant legal discourse surrounding copyright protections. Two primary issues
stand out: the risk of copyright infringement and the difficulty in obtaining copyright
protection for AI-generated content [35].
At the USA level, the U.S. Copyright Office maintains a firm stance that copyright law
protects only works created with human authorship. This principle was emphasized in the
2025 Report on Copyright and Artificial Intelligence, which clarifies that works generated
entirely by AI, devoid of meaningful human intervention, do not qualify for copyright
protection [40][47]. Consequently, code produced solely by AI cannot be registered for
copyright, placing it in a category that may be freely used by anyone unless protected under
another legal framework, such as trade secrets [32][47]. However, when human developers
engage in structuring, selecting, and refining AI-assisted outputs, they can establish
sufficient human authorship to secure copyright for those contributions [40][47]. This
distinction is critical for businesses that integrate AI tools into their development processes,
as it underscores the importance of human involvement in obtaining copyright protections.

Uses of Generative AI Tools

Generative AI tools are utilized in multiple ways within the realm of software development.
• Generating and Reviewing Artifacts: AI tools facilitate the creation and refinement
of project documents, including requirements specifications and design documents,
ensuring accuracy and completeness in project deliverables [5].
• Supporting Ideation Processes: AI assists in brainstorming sessions and design
thinking by generating novel ideas and concepts, thereby fostering creativity during
the development phase [5].
• Resolving Doubts in Code Construction: Developers can leverage AI to
troubleshoot and resolve technical issues encountered during programming,
enhancing problem-solving capabilities [5].

Key Features of Generative AI Tools

Pag. 309 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Generative AI tools are equipped with advanced functionalities designed to aid developers
in their coding tasks.
• Code Completion and Generation: Tools like GitHub Copilot, ChatGPT and
Amazon CodeWhisperer enable developers to write code more efficiently by
suggesting code completions and even entire functions based on their coding style
and intent [19].
• Language Understanding: These tools utilize sophisticated natural language
processing algorithms to interpret the intent behind code, allowing developers to
work across various programming languages without hindrance [19].
• Integration with IDEs: Many generative AI tools seamlessly integrate with popular
integrated development environments (IDEs) such as Visual Studio Code and
IntelliJ, providing developers with immediate access to AI capabilities without
disrupting their workflow [19].

Challenges of Generative AI Tools

Despite the advantages, the integration of generative AI tools in software development also
presents several challenges. These include issues related to explainability and
interpretability of AI decisions, as well as concerns regarding data governance and the
quality of training data used in these models [9][20]. Addressing these challenges requires
a comprehensive approach that emphasizes transparency and ethical guidelines in the use
of generative AI technologies [12].

3.1 Integration of AI Generative Tools with Java

Overview of AI in Java Development
The integration of AI generative tools in Java programming is reshaping the software
development landscape, enhancing productivity, and enabling developers to focus on more
complex tasks. Generative AI tools such as Co-Pilot and Codex can generate Java code
from natural language descriptions, significantly reducing the time developers spend on
producing boilerplate code, thus improving overall efficiency and output quality [6][21].
As Java remains a leading platform for AI projects due to its flexibility, comprehensive
libraries, and large-scale data processing capabilities, the collaboration between AI and Java
is becoming increasingly vital [7][22].
At this moment, we can conclude that AI tools for code generation have significantly
transformed the software development landscape, offering a new approach to writing and
optimizing code. These intelligent systems leverage advanced technologies such as
Artificial Intelligence (AI) and Machine Learning (ML) to assist developers by generating
code based on natural language descriptions and existing code contexts.
Pag. 310 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Enhancements in Developer Productivity

Participants in studies have reported a positive correlation between the use of generative AI
tools and their perceived productivity. These tools not only streamline workflows by
consolidating multiple functions into a single application but also provide quick access to
information that enhances team communication and collaboration [5]. Despite challenges
related to reliability and the need for refinement of generated outputs, developers noted that
generative AI tools facilitate gains in efficiency and flow, allowing them to produce relevant
content such as code, reports, and design models more effectively [5][8].
Challenges in Integration
While generative AI tools present numerous advantages, developers also face challenges
such as ensuring the reliability of generated outputs and the need for precise prompting to
achieve accurate results. The absence of sources to validate the information provided by
these tools further complicates their integration into sensitive projects [5][21]. Moreover,
there are security concerns that can restrict the use of generative AI tools when dealing with
sensitive data in software development contexts [5].
Future Directions and Best Practices
Looking ahead, the Java community is actively embracing AI-driven solutions, with a high
percentage of developers expressing interest in building AI-powered applications using
Java [22]. As Java developers prepare for the future, they emphasize the importance of
essential elements like Retrieval-Augmented Generation (RAG) and embeddings for
advanced knowledge retrieval [23][22]. These approaches will be crucial in developing
practical, enterprise-ready features that are reliable and secure, enabling developers to
navigate complex decision-making scenarios effectively.
How AI Code Generation Works
At its core, AI code generation involves using sophisticated algorithms that are trained on
extensive datasets sourced from publicly available code repositories. Large Language
Models (LLMs) employ Deep Learning (DL) techniques to understand programming
patterns and best practices, enabling them to suggest or create code snippets that meet user-
defined functionalities [42][43]. Developers input plain text prompts that outline their
coding requirements, and the AI tools respond with relevant code suggestions or full
functions, streamlining the development process and reducing manual effort [44][45].
Current AI code generation tools are capable of producing code in various programming
languages based on natural language input. Developers can articulate the desired
functionality, and the AI translates these instructions into contextually appropriate code,
enhancing productivity and accessibility for programmers of all skill levels [42][44]. As
already mentioned, notable tools in this domain include GitHub Copilot, Amazon Code
Pag. 311 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Whisperer or ChatGPT, which offer features such as code completion, snippets, and even
entire functions to assist in the coding workflow [45][31].
In addition to generating code, AI tools also conduct automatic reviews of existing
codebases to identify security vulnerabilities, bugs, and common coding errors. This
process not only helps maintain code quality but also allows developers to save time by
automating repetitive tasks that would otherwise require significant manual intervention
[42][43]. However, it is important to note that the code generated by these tools should still
undergo rigorous review by human developers to mitigate the risks associated with
inaccuracies and security vulnerabilities [44][45].

4. Case Studies and Examples

Pilot Case Study on Generative AI in Software Development
A pilot case study was conducted to investigate the impact of generative AI tools on
software development productivity. The study focused on a single company and involved
various software development roles, including developers, quality assurance (QA)
professionals, and designers. The primary objectives were to explore how different
professionals utilize generative AI tools in their tasks and to understand the tools' perceived
effects on productivity across diverse project configurations and methodologies [5].

Methodology and Data Collection

The research utilized a case study methodology that involved questionnaires with open-
ended questions and observations. The questionnaires aimed to gather insights from
participants who volunteered to use generative AI tools, ensuring minimal disruption to
their work routines [5]. Observational data from company communication channels, such
as Slack, were also collected to identify potential participants and assess discussions
surrounding generative AI [5].

Findings on Perceived Productivity

Participants reported a positive effect of generative AI tools on their perceived productivity,
particularly in terms of efficiency gains across their software development activities. The
tools facilitated time optimization by consolidating multiple individual tools into a single
streamlined workflow. Despite facing some challenges, such as concerns over reliability,
most software professionals acknowledged improvements in their ability to create relevant
and insightful outputs, including code, design models, and documentation [5].

Pag. 312 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Indirect Impacts on Communication and Collaboration

While not explicitly stated by the participants, the use of generative AI tools appeared to
indirectly enhance communication and collaboration within development teams. By
providing quick access to information and facilitating knowledge sharing, these tools
allowed team members to align their understanding and work towards common goals more
effectively [5].

Broader Implications and Applications of Generative AI Tools

Generative AI tools offer a variety of applications that can significantly aid software
developers. These applications include refactoring and code improvement, documentation
generation, learning assistance, prototype development, and data transformation and
analysis [24]. For example, developers can use AI to automate data cleansing tasks or
rapidly develop proof-of-concept applications, thereby reducing the upfront investment of
time and resources [24]. Bogdan Mykhaylovych, Technical Director at Softjourn,
emphasized the importance of integrating tailored AI models into specific business domains
to further enhance productivity for developers and product owners [24]. This integration
highlights the adaptability of generative AI tools across various aspects of the software
development lifecycle, reinforcing their potential to redefine traditional programming
methodologies.

Ethical Considerations
The integration of AI generative tools into Java programming raises several ethical
concerns, particularly regarding ownership and originality in AI-generated code. Issues
related to intellectual property rights and copyright are increasingly becoming significant
in discussions about the ethical use of such technologies [11].

Algorithmic Bias
One of the most pressing ethical considerations in machine learning and AI is algorithmic
bias, which refers to systematic errors that lead to discrimination against certain groups
based on the outputs of a program [10][25]. This bias can arise when training datasets are
unrepresentative, leading the AI to produce results that favor specific demographics over
others [26]. For instance, algorithms may exhibit racial bias in healthcare risk assessments
or show discrimination in hiring practices, as evidenced by cases involving biased
recruitment tools [27].

Mitigating Bias
Pag. 313 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

To address algorithmic bias, developers and operators of AI algorithms are urged to

implement best practices at various stages of the AI system's lifecycle [28]. This includes
examining the training data for representation and ensuring that the model does not
perpetuate existing inequalities. Additionally, efforts such as using AI Fairness 360,
Fairlearn, and other fairness tools can help identify and reduce bias in AI systems [27].

The Role of Policymakers

Policymakers also play a critical role in addressing these issues. The establishment of
regulatory sandboxes could provide a framework for testing AI technologies while
developing appropriate regulations to curb biases [13]. Such environments allow for
innovation in technology alongside regulatory evolution, helping ensure that ethical
standards are upheld without stifling technological advancement.

Broader Ethical Implications

Beyond technical concerns, the ethical implications of AI-generated tools extend to their
societal impacts. Developers must consider how their models may influence individuals
and communities, particularly in high-stakes areas like job applicant suitability assessments
[11]. Discussions surrounding the ethical framework of machine learning emphasize the
need for guardrails to prevent the amplification of systemic discrimination and other
negative consequences [13].

5. Copyright Issues for Java Code Generated with AI Tools

The copyright implications of Java code generated by artificial intelligence (AI) tools have
sparked significant legal debate as the integration of these technologies into software
development raises questions about authorship and intellectual property rights. Historically,
U.S. copyright law mandates that only works created by human beings are eligible for
protection, a principle reaffirmed by the U.S. Copyright Office and recent court rulings such
as Thaler v. Perlmutter, which ruled that AI-generated works without meaningful human
involvement cannot be copyrighted [31][32][33].
As a result, developers and companies face considerable risks when utilizing AI-generated
code, as such outputs may not only be unprotected by copyright but could also infringe on
existing copyrighted works. The uniqueness of the situation is compounded by the nature
of AI code generation tools, which rely on vast datasets that may include copyrighted
material. This raises the potential for legal conflicts when companies modify or deploy AI-
generated code, as they could inadvertently violate the rights of original copyright holders
[34][32].
Pag. 314 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Legal experts emphasize that businesses must navigate these complexities carefully,
documenting their creative processes and consulting with legal professionals to mitigate
infringement risks and ensure compliance with copyright law [35][36]. Moreover, the
evolving legal landscape surrounding AI-generated content necessitates ongoing discourse
on the definition of authorship and the applicability of copyright in the digital age. Current
legislative efforts and high-profile lawsuits are beginning to address these challenges, but
many uncertainties remain. The juxtaposition of rapid technological advancement with
existing copyright frameworks highlights the need for reform to address the implications of
generative AI on intellectual property rights, particularly in the realm of software
development [37][38][39].

The use of AI in generating code also introduces substantial copyright infringement risks.
As copyright automatically applies to original source code, modifying existing works with
AI could infringe on the exclusive rights held by the original copyright owner [34].
Businesses must seek permission from copyright holders when using AI to modify protected
works, much like any traditional method of content alteration. Additionally, the landscape
of copyright law necessitates that organizations document their independent creation
processes and consult legal professionals about copyright risks before reusing code. This
proactive approach is essential given the complex nature of fair use determinations and the
legal ramifications of infringement claims [36]. The current legal framework, in all
countries, requires businesses to critically assess their reliance on AI in creative fields and
to navigate the copyright landscape with very much care. While the US Copyright Office
and other similar bodies worldwide have yet to provide definitive guidance on the future of
copyright in relation to AI-generated works, it has been made clear that human authorship
remains a cornerstone of copyright protection [32].
The legal landscape regarding the copyright of AI-generated works, particularly Java code,
is evolving as courts address the implications of artificial intelligence on existing copyright
law. As of now, the legality of using copyrighted content to train AI models remains
unsettled, with outcomes heavily reliant on jurisdiction and specific case circumstances
[37][48]. This uncertainty has led to a growing demand for legislative or regulatory
solutions to clarify when training on copyrighted content is permissible [37]. This places
an added responsibility on developers and companies to ensure that their use of AI tools
complies with copyright law to mitigate potential legal and financial repercussions. The
U.S. Copyright Office has released reports addressing the copyrightability of outputs
created using generative AI, affirming that existing copyright principles can accommodate
these new technologies. The Office maintains that AI outputs may be copyrightable when a
human author contributes significant creative elements, such as perceptible modifications
or creative arrangements, rather than merely providing prompts [41][50]. This stance sets a
high bar for the recognition of AI-generated works under copyright law, leaving many
Pag. 315 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

boundaries regarding protectable and unprotectable works still undefined [50]. Courts have
also pointed out that AI-generated outputs could infringe copyright if they closely resemble
existing works, raising significant legal questions regarding originality and authorship
[49][50].

As AI tools continue to transform coding practices, their legal and ethical ramifications
warrant thorough examination. The lack of a clear path to copyright protection for AI-
generated Java code emphasizes the importance of human involvement in the creative
process to secure intellectual property rights. The outcome of ongoing legal battles and
potential legislative changes will be pivotal in shaping the future of copyright in relation to
AI technologies, ultimately influencing how developers and businesses engage with these
innovative tools [40][37][41].

5.1 The Impact of Generative AI on Copyright

The advent of generative artificial intelligence (AI) technologies has raised new challenges
for traditional notions of authorship in copyright law. Current interpretations of copyright
require that an original work of authorship must be created by a human being, a stance
upheld by the U.S. Copyright Office [32]. Recent court rulings in the USA, such as Thaler
v. Perlmutter, have reinforced this position, stating that works generated by AI without
human intervention do not qualify for copyright protection [33]. This limitation has
significant implications for the use of AI in creative domains, particularly when businesses
aim to utilize AI-generated content, as they may not enjoy the typical intellectual property
protections if such content is challenged by rivals [33].

5.2 Current Legislative Landscape and Ongoing Legal Battles

The legislative landscape concerning copyright and AI-generated Java code is rapidly
evolving as governments worldwide grapple with the implications of generative AI
technologies. In March 2024, the European Parliament approved the EU AI Act, which
mandates that providers of general-purpose AI models comply with existing copyright laws,
particularly regarding the reservation of rights for training data [39]. This act signifies a
proactive approach by the EU in addressing copyright concerns related to AI, emphasizing
the need for transparency and compliance with copyright norms. In the USA, various states
have introduced their own legislation targeting generative AI, with notable bills emerging
in Colorado and California in 2024. These state-level initiatives reflect a recognition of the
need to establish a regulatory framework that balances innovation with the protection of
intellectual property rights [39]. However, the U.S. federal stance remains somewhat

Pag. 316 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ambiguous, with courts and legislators still deliberating on the copyrightability of AI-
generated content [53].
The approaches to copyright for AI-generated works differ significantly among
jurisdictions. In the U.S., the Copyright Office maintains a stringent requirement for human
authorship, stating that copyright entitlement can only be established if the work originates
from a human agent [33]. This is particularly relevant to Java code generated by AI tools,
as the lack of a human author would preclude such code from receiving copyright protection
under current U.S. law [34]. Conversely, the EU has initiated frameworks that could
potentially allow for broader interpretations of AI-generated works. The recent introduction
of exceptions for text and data mining (TDM) indicates a willingness to adapt existing laws
to facilitate AI development while safeguarding creator rights [37]. This distinction
highlights a significant divergence from U.S. legislation, where the fair use doctrine
provides a more flexible, albeit less clear-cut, framework for utilizing copyrighted material
for AI training.
Experts advocate for international harmonization of copyright laws concerning AI,
proposing common principles that could facilitate a global approach to managing AI-
generated content [37]. Such harmonization could involve agreements that permit the use
of works for AI training under specific conditions, thereby balancing the interests of
creators with the need to foster innovation in AI technologies. One suggestion includes
establishing a global data licensing framework that allows AI firms to contribute to a fund
distributing payments to content creators in exchange for licenses to train on published
content [37]. The ongoing discussions around AI and copyright reflect a critical juncture
for policymakers. Achieving a consensus that aligns with technological advancements
while protecting the rights of creators will likely require extensive international dialogue,
possibly facilitated through organizations like the World Intellectual Property Organization
(WIPO) or within trade agreements. The outcome of these discussions will be pivotal in
shaping the future of copyright as it relates to AI-generated Java code and other creative
works.

As AI continues to advance, high-profile lawsuits have emerged against companies like

OpenAI and Stability AI, with plaintiffs alleging unauthorized use of copyrighted material
to train AI models. The ongoing case of New York Times Co. v. Microsoft serves as a critical
example of how courts are grappling with these issues, setting a precedent for the
intersection of copyright law and AI technology. These cases could have far-reaching
implications for the future of copyright and AI-generated content, as the outcomes may
influence both the interpretation of fair use and the rights of creators in the digital age
[38][51][52].

Pag. 317 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

5.3 Implications for Java Code Generated with AI

When considering Java code generated by AI tools, the existing legal framework suggests
that if the code lacks direct human authorship, it may not be eligible for copyright
protection. This situation necessitates that developers and businesses exercise caution, as
using or modifying AI-generated code could potentially infringe upon the copyright rights
of existing works, unless proper permissions are obtained from the original copyright
holders [34]. As the legal landscape continues to evolve alongside advancements in AI
technologies, ongoing discourse regarding the definitions of authorship and the
applicability of copyright law will be crucial in shaping the future of intellectual property
rights in the realm of software development and beyond.
Looking ahead, it is anticipated that a more settled equilibrium will emerge between AI
developers, content creators, and consumers. This will likely involve clearer guidelines on
permissible data usage for AI training, ensuring that creators can benefit from AI utilization
without facing undue harm. Additionally, it is crucial to foster an environment where
diverse creative content continues to thrive, balancing innovation with the protection of
human authorship [37]. Ultimately, while challenges exist, the potential for AI to enhance
programming and creativity remains substantial, provided that stakeholders actively engage
with the legal and ethical implications [19].

6. Conclusions and Future Trends

The landscape of Java programming with AI generative tools is poised for significant
evolution in the coming years. As the integration of AI technologies into software
development becomes more prevalent, we can expect several transformative trends to
emerge.

Enhanced Collaboration Between AI and Developers

Rather than viewing AI as a competitor, the future of programming will increasingly focus
on collaboration. Developers will leverage AI tools to automate mundane tasks such as code
generation, debugging, and testing, allowing them to concentrate on more complex
problem-solving aspects of software development [15][29]. This symbiotic relationship is
anticipated to boost productivity, with AI tools offering developers real-time information
and support during coding [8].

Expanding Use Cases for AI in Java Development

AI's capabilities will likely extend into more specialized areas of Java development. Use
cases such as natural language processing (NLP) and dynamic pricing systems will benefit
Pag. 318 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

from advanced AI models embedded in Java applications, enhancing their functionality and
efficiency [22]. The integration of AI frameworks into existing systems will also enable
developers to implement machine learning models more effectively, addressing complex
challenges like customer churn prediction and recommendation systems.

Addressing Ethical and Security Challenges

As AI becomes integral to coding practices, addressing ethical and security concerns will
be paramount. Developers will need to navigate potential pitfalls related to security
vulnerabilities, intellectual property issues, and code quality management [14]. Robust risk
management strategies will be essential as organizations seek to balance the advantages of
AI tools with the associated risks.

Continuous Learning and Adaptation

The rapid pace of AI development necessitates continuous learning and adaptability among
programmers. Professionals will need to stay updated on evolving AI technologies and
methodologies to maximize their effectiveness in collaboration with AI tools [16][29].
Educational institutions and training programs are likely to adapt their curricula to
incorporate AI literacy, equipping future developers with the necessary skills to thrive in
this changing environment.

Acknowledgment
The paper is based on research carried out in part within the Center for Computational
Science and Machine Intelligence (CSMI) of the Romanian-American University’s School
of Computer Science for Business Management, and the Center for Human Rights and
Migration (CDOM) of the Romanian-American University’s School of Law.

References
[1] Java (programming language) – Wikipedia;
https://en.wikipedia.org/wiki/Java_(programming_language); last access: May 04, 2025
[2] Java Features; March 24, 2025; https://www.geeksforgeeks.org/java-features/; last
access: May 04, 2025

Pag. 319 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[3] How Java Is Different From Other Programming Languages; https://www.fita.in/how-

java-is-different-from-other-programming-languages/; last access: May 04, 2025
[4] Bello Adeshola, Amorighoye Otse; Java Programming Language;
https://amorserv.com/insights/java-programming-language
[5] Kumar Vinod; Exploring the Key Features of Java Programming Language;
https://medium.com/@vinodkumarbheel61/exploring-the-key-features-of-java-
programming-language-315f92e58869
[6] Bloch-Wehba Hannah; Transparency's AI Problem; June 17, 2021;
https://knightcolumbia.org/content/transparencys-ai-problem
[7] Coutinho Mariana, Marques Lorena, Santos Anderson, Dahia Marcio, Franca Cesar,
de Souza Santos Ronnie; The Role of Generative AI in Software Development
Productivity: A Pilot Case Study; Proceedings of the 32nd ACM Symposium on the
Foundations of Software Engineering (FSE ’24), November 15–19, 2024, Porto de
Galinhas, Brazil; https://arxiv.org/html/2406.00560v1
[8] Generative AI Tools for Software Development | Coding Temple;
https://www.codingtemple.com/blog/generative-ai-tools-for-software-development/; last
access: May 04, 2025
[9] AI Transparency: Fundamental pillar for ethical and safe AI; April 08, 2025;
https://www.plainconcepts.com/ai-transparency/; last access: May 04, 2025
[10] Hassani Hesam Sheikh; AI Governance: Frameworks, Tools, Best Practices;
September 04, 2024; https://www.datacamp.com/blog/ai-governance
[11] Ruiz Julie Tyler; Generative AI Ethics: AI Risks, Benefits, and Best Practices;
October 27, 2024; https://www.coursera.org/articles/generative-ai-ethics
[12] Laforge Guillaume; Generative AI text generation in Java with PaLM and
LangChain4J; December 04, 2023;
https://codelabs.developers.google.com/codelabs/genai-text-gen-java-palm-langchain4j
[13] The Impact of Generative AI on Java Development: From Code to Deployment;
November 11, 2024; https://medium.com/@speckyfox.technologies/the-impact-of-
generative-ai-on-java-development-from-code-to-deployment-8d1e3d2a35d8; last access:
May 04, 2025
[14] The Future of Java Development: 13 AI Tools Transforming Productivity; January
15; 2025; https://groovetechnology.com/blog/technologies/ai-for-java-developers/; last
access: May 04, 2025
[15] Sakovich Natallia; How to Implement AI in Java: A Step-by-Step Guide; May 20,
2025 https://sam-solutions.com/blog/how-to-implement-ai-in-java/

Pag. 320 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[16] The Power of AI-Assisted Coding: Streamlining Rapid Development in Java;

September 11, 2024; https://blog.kodezi.com/the-power-of-ai-assisted-coding-
streamlining-rapid-development-in-java/; last access: May 04, 2025
[17] Selvasingh Asir; The State of Coding the Future with Java and AI; May 12, 2025;
https://devblogs.microsoft.com/java/the-state-of-coding-the-future-with-java-and-ai/
[18] Neville Meghan; Can AI Really Code? Understanding the Strengths and Limitations
of AI Coding Tools; April 25, 2025; https://softjourn.com/insights/ai-coding-tools
[19] Nasimov Salmany; Taming the code generation beast — How responsible is your AI
adoption in Java; October 29, 2024; https://digma.ai/taming-the-code-generation-beast-
how-responsible-is-your-ai-adoption-in-java/
[20] Chen, Z.; Ethics and discrimination in artificial intelligence-enabled recruitment
practices. Humanit Soc Sci Commun 10, 567 (2023). https://doi.org/10.1057/s41599-023-
02079-x; https://www.nature.com/articles/s41599-023-02079-x
[21] Wood Maria; What Are the Risks of Algorithmic Bias in Higher Education?; May
2021; https://www.everylearnereverywhere.org/blog/what-are-the-risks-of-algorithmic-
bias-in-higher-education/
[22] Levin Cait; AI Prompt Engineering: How to Instruct AI Models to Avoid AI Bias and
Get The Results You Want; September 4, 2024; https://www.glideapps.com/blog/ai-
prompt-engineering
[23] Sinha Swarnabha; How to Reduce Bias in AI? - Scaler Topics; December 8, 2023;
https://www.scaler.com/topics/artificial-intelligence-tutorial/ai-bias/
[24] Awan Abid Ali;l What is Algorithmic Bias?; July 17, 2023;
https://www.datacamp.com/blog/what-is-algorithmic-bias
[25] Turner Lee Nicol, Resnick Paul, and Barton Genie; Algorithmic bias detection and
mitigation: Best practices and policies to reduce consumer harms; May 22, 2019;
https://www.brookings.edu/articles/algorithmic-bias-detection-and-mitigation-best-
practices-and-policies-to-reduce-consumer-harms/
[26] Siedsma Andrea; Will AI Replace Programmers? Navigating the Future of Coding;
March 22, 2024; https://extendedstudies.ucsd.edu/news-events/extended-studies-
blog/will-ai-replace-programmers-navigating-the-future-of-coding
[27] Jasińska Dorota, Puczyk Andrzej; Top Generative AI Platforms and Tools for
Software Development; May 07, 2025; https://neontri.com/blog/generative-ai-platforms/
[28] Zaman Sian; December 29. 2024; Risks Of Using AI In Software Development - Is It
All Bad?; https://impalaintech.com/blog/risks-of-ai-software-development/

Pag. 321 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[29] Zeichick Alan; 7 Ways GenAI Can Help Improve Software Development; October 8,
2024; https://www.oracle.com/artificial-intelligence/generative-ai/generative-ai-software-
development/
[30] 10 Best Java IDE For Developers in 2025; December 13, 2024;
https://www.geeksforgeeks.org/best-java-ide-for-developers/; last access: May 04, 2025
[31] Caldwell Mackenzie; What Is an “Author”?-Copyright Authorship of AI Art Through
a Philosophical Lens; Houston Law Review, Vol. 61, Issue 2, 2023;
https://houstonlawreview.org/article/92132-what-is-an-author-copyright-authorship-of-ai-
art-through-a-philosophical-lens
[32] Artificial Intelligence; January 30, 2025;
https://libguides.washburn.edu/c.php?g=1377087&p=10201459/; last access: May 04,
2025
[33] Laney Douglas B.; Copyright or Copywrong? AI’s Intellectual Property Conundrum;
Forbes Contributors, February 11, 2025;
https://www.forbes.com/sites/douglaslaney/2025/02/11/copyright-or-copywrong-ais-
intellectual-property-paradox/
[34] Vana Jim; Court Rules AI Can’t Author a Copyrighted Work; March 19, 2025;
https://www.schwabe.com/publication/court-rules-ai-cant-author-a-copyrighted-work/
[35] What is AI code generation?; https://aws.amazon.com/what-is/ai-coding/; last access:
May 04, 2025
[36] James Kay; How to Use AI Code Generation to Enhance Developer Productivity;
https://www.getambassador.io/blog/ai-code-generation-boost-developer-productivity
[37] AI code generation; https://www.sonarsource.com/learn/ai-code-generation/; last
access: May 04, 2025
[38] AI Code Generation: The Risks and Benefits of AI in Software;
https://www.legitsecurity.com/aspm-knowledge-base/ai-code-generation-benefits-and-
risks; last access: May 04, 2025
[39] Artificial intelligence and copyright;
https://en.wikipedia.org/wiki/Artificial_intelligence_and_copyright; last access: May 04,
2025
[40] Pelkonen Tuomas; Legal Insights into AI-Driven Game Development; October 16,
2024; https://nordialaw.com/legal-insights-into-ai-driven-game-development/
[41] Mathur Atreya; Recent Developments in AI, Art & Copyright: Copyright Office
Report & New Registrations; March 4, 2025; https://itsartlaw.org/2025/03/04/recent-
developments-in-ai-art-copyright-copyright-office-report-new-registrations/

Pag. 322 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[42] Furdock Brandon; Navigating the Legal Landscape of AI-Generated Code:

Ownership and Liability Challenges; May 19, 2025;
https://www.mbhb.com/intelligence/snippets/navigating-the-legal-landscape-of-ai-
generated-code-ownership-and-liability-challenges/
[43] Sumner Michael; Protecting Source Code from Copyright Infringement: A Guide;
February 09, 2025; https://www.scoredetect.com/blog/posts/protecting-source-code-from-
copyright-infringement-a-guide
[44] Masood Adnan; Intellectual Property Rights and AI-Generated Content — Issues in
Human Authorship, Fair Use Doctrine, and Output Liability; April 04, 2025;
https://medium.com/@adnanmasood/intellectual-property-rights-and-ai-generated-
content-issues-in-human-authorship-fair-use-8c7ec9d6fdc3
[45] Gewirtz David; If ChatGPT produces AI-generated code for your app, who does it
really belong to?; December 21, 2024; https://www.zdnet.com/article/if-chatgpt-produces-
ai-generated-code-for-your-app-who-does-it-really-belong-to/
[46] Zirpoli Christopher T.; Generative Artificial Intelligence and Copyright Law;
September 29, 2023; https://www.congress.gov/crs-product/LSB10922
[47] Oratz Lisa T.; Copyright Office Solidifies Stance on the Copyrightability of AI-
Generated Works; March 19, 2025; https://perkinscoie.com/insights/update/copyright-
office-solidifies-stance-copyrightability-ai-generated-works
[48] Copyright Office Releases Part 2 of Artificial Intelligence Report;
https://www.copyright.gov/newsnet/2025/1060.html; last access: May 04, 2025
[49] Artificial Intelligence and the Creative Double Bind - Chapter 2; Harvard Law
Review; Vol. 138, Issue 6, April 2025; https://harvardlawreview.org/print/vol-
138/artificial-intelligence-and-the-creative-double-bind/
[50] Klosek Katherine, Blumenthal Marjory S.; Training Generative AI Models on
Copyrighted Works Is Fair Use; January 23, 2025; https://www.arl.org/blog/training-
generative-ai-models-on-copyrighted-works-is-fair-use/
[51] Kim Rachel; AI and Copyright in 2023: In the Courts; January 4, 2024;
https://copyrightalliance.org/ai-copyright-courts/
[52] Stalnaker Trevor, Wintersgill Nathan, Chaparro Oscar, Heymann Laura A., Di Penta
Massimiliano, German Daniel M., Poshyvanyk Denys; Developer Perspectives on
Licensing and Copyright Issues Arising from Generative AI for Coding; November 2024;
https://arxiv.org/html/2411.10877v1
[53] Bondari Negar; AI, Copyright, and the Law: The Ongoing Battle Over Intellectual
Property Rights; February 04, 2025; https://sites.usc.edu/iptls/2025/02/04/ai-copyright-
and-the-law-the-ongoing-battle-over-intellectual-property-rights/
Pag. 323 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Bibliography
Awan Abid Ali;l What is Algorithmic Bias?; July 17, 2023;
https://www.datacamp.com/blog/what-is-algorithmic-bias
Bello Adeshola, Amorighoye Otse; Java Programming Language;
https://amorserv.com/insights/java-programming-language
Bloch-Wehba Hannah; Transparency's AI Problem; June 17, 2021;
https://knightcolumbia.org/content/transparencys-ai-problem
Bondari Negar; AI, Copyright, and the Law: The Ongoing Battle Over Intellectual
Property Rights; February 04, 2025; https://sites.usc.edu/iptls/2025/02/04/ai-copyright-
and-the-law-the-ongoing-battle-over-intellectual-property-rights/
Caldwell Mackenzie; What Is an “Author”?-Copyright Authorship of AI Art Through a
Philosophical Lens; Houston Law Review, Vol. 61, Issue 2, 2023;
https://houstonlawreview.org/article/92132-what-is-an-author-copyright-authorship-of-ai-
art-through-a-philosophical-lens
Chen, Z.; Ethics and discrimination in artificial intelligence-enabled recruitment practices.
Humanit Soc Sci Commun 10, 567 (2023). https://doi.org/10.1057/s41599-023-02079-x;
https://www.nature.com/articles/s41599-023-02079-x
Coutinho Mariana, Marques Lorena, Santos Anderson, Dahia Marcio, Franca Cesar, de
Souza Santos Ronnie; The Role of Generative AI in Software Development Productivity:
A Pilot Case Study; Proceedings of the 32nd ACM Symposium on the Foundations of
Software Engineering (FSE ’24), November 15–19, 2024, Porto de Galinhas, Brazil;
https://arxiv.org/html/2406.00560v1
Furdock Brandon; Navigating the Legal Landscape of AI-Generated Code: Ownership
and Liability Challenges; May 19, 2025;
https://www.mbhb.com/intelligence/snippets/navigating-the-legal-landscape-of-ai-
generated-code-ownership-and-liability-challenges/
Gewirtz David; If ChatGPT produces AI-generated code for your app, who does it really
belong to?; December 21, 2024; https://www.zdnet.com/article/if-chatgpt-produces-ai-
generated-code-for-your-app-who-does-it-really-belong-to/
Hassani Hesam Sheikh; AI Governance: Frameworks, Tools, Best Practices; September
04, 2024; https://www.datacamp.com/blog/ai-governance
James Kay; How to Use AI Code Generation to Enhance Developer Productivity;
https://www.getambassador.io/blog/ai-code-generation-boost-developer-productivity
Pag. 324 / 350
Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Jasińska Dorota, Puczyk Andrzej; Top Generative AI Platforms and Tools for Software
Development; May 07, 2025; https://neontri.com/blog/generative-ai-platforms/
Kim Rachel; AI and Copyright in 2023: In the Courts; January 4, 2024;
https://copyrightalliance.org/ai-copyright-courts/
Klosek Katherine, Blumenthal Marjory S.; Training Generative AI Models on
Copyrighted Works Is Fair Use; January 23, 2025; https://www.arl.org/blog/training-
generative-ai-models-on-copyrighted-works-is-fair-use/
Kumar Vinod; Exploring the Key Features of Java Programming Language;
https://medium.com/@vinodkumarbheel61/exploring-the-key-features-of-java-
programming-language-315f92e58869
Laforge Guillaume; Generative AI text generation in Java with PaLM and LangChain4J;
December 04, 2023; https://codelabs.developers.google.com/codelabs/genai-text-gen-
java-palm-langchain4j
Laney Douglas B.; Copyright Or Copywrong? AI’s Intellectual Property Conundrum;
Forbes Contributors, February 11, 2025;
https://www.forbes.com/sites/douglaslaney/2025/02/11/copyright-or-copywrong-ais-
intellectual-property-paradox/
Levin Cait; AI Prompt Engineering: How to Instruct AI Models to Avoid AI Bias and Get
The Results You Want; September 4, 2024; https://www.glideapps.com/blog/ai-prompt-
engineering
Masood Adnan; Intellectual Property Rights and AI-Generated Content — Issues in
Human Authorship, Fair Use Doctrine, and Output Liability; April 04, 2025;
https://medium.com/@adnanmasood/intellectual-property-rights-and-ai-generated-
content-issues-in-human-authorship-fair-use-8c7ec9d6fdc3
Mathur Atreya; Recent Developments in AI, Art & Copyright: Copyright Office Report &
New Registrations; March 4, 2025; https://itsartlaw.org/2025/03/04/recent-developments-
in-ai-art-copyright-copyright-office-report-new-registrations/
Nasimov Salmany; Taming the code generation beast — How responsible is your AI
adoption in Java; October 29, 2024; https://digma.ai/taming-the-code-generation-beast-
how-responsible-is-your-ai-adoption-in-java/
Neville Meghan; Can AI Really Code? Understanding the Strengths and Limitations of AI
Coding Tools; April 25, 2025; https://softjourn.com/insights/ai-coding-tools
Oratz Lisa T.; Copyright Office Solidifies Stance on the Copyrightability of AI-Generated
Works; March 19, 2025; https://perkinscoie.com/insights/update/copyright-office-
solidifies-stance-copyrightability-ai-generated-works

Pag. 325 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Pelkonen Tuomas; Legal Insights into AI-Driven Game Development; October 16, 2024;
https://nordialaw.com/legal-insights-into-ai-driven-game-development/
Ruiz Julie Tyler; Generative AI Ethics: AI Risks, Benefits, and Best Practices; October 27,
2024; https://www.coursera.org/articles/generative-ai-ethics
Sakovich Natallia; How to Implement AI in Java: A Step-by-Step Guide; May 20, 2025
https://sam-solutions.com/blog/how-to-implement-ai-in-java/
Selvasingh Asir; The State of Coding the Future with Java and AI; May 12, 2025;
https://devblogs.microsoft.com/java/the-state-of-coding-the-future-with-java-and-ai/
Siedsma Andrea; Will AI Replace Programmers? Navigating the Future of Coding; March
22, 2024; https://extendedstudies.ucsd.edu/news-events/extended-studies-blog/will-ai-
replace-programmers-navigating-the-future-of-coding
Sinha Swarnabha; How to Reduce Bias in AI? - Scaler Topics; December 8, 2023;
https://www.scaler.com/topics/artificial-intelligence-tutorial/ai-bias/
Sumner Michael; Protecting Source Code from Copyright Infringement: A Guide;
February 09, 2025; https://www.scoredetect.com/blog/posts/protecting-source-code-from-
copyright-infringement-a-guide
Stalnaker Trevor, Wintersgill Nathan, Chaparro Oscar, Heymann Laura A., Di Penta
Massimiliano, German Daniel M., Poshyvanyk Denys; Developer Perspectives on
Licensing and Copyright Issues Arising from Generative AI for Coding; November 2024;
https://arxiv.org/html/2411.10877v1
Turner Lee Nicol, Resnick Paul, and Barton Genie; Algorithmic bias detection and
mitigation: Best practices and policies to reduce consumer harms; May 22, 2019;
https://www.brookings.edu/articles/algorithmic-bias-detection-and-mitigation-best-
practices-and-policies-to-reduce-consumer-harms/
Vana Jim; Court Rules AI Can’t Author a Copyrighted Work; March 19, 2025;
https://www.schwabe.com/publication/court-rules-ai-cant-author-a-copyrighted-work/
Wood Maria; What Are the Risks of Algorithmic Bias in Higher Education?; May 2021;
https://www.everylearnereverywhere.org/blog/what-are-the-risks-of-algorithmic-bias-in-
higher-education/
Zaman Sian; December 29. 2024; Risks Of Using AI In Software Development - Is It All
Bad?; https://impalaintech.com/blog/risks-of-ai-software-development/
Zeichick Alan; 7 Ways GenAI Can Help Improve Software Development; October 8,
2024; https://www.oracle.com/artificial-intelligence/generative-ai/generative-ai-software-
development/

Pag. 326 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Zirpoli Christopher T.; Generative Artificial Intelligence and Copyright Law; September
29, 2023; https://www.congress.gov/crs-product/LSB10922

*** 10 Best Java IDE For Developers in 2025; December 13, 2024;
https://www.geeksforgeeks.org/best-java-ide-for-developers/; last access: May 04, 2025
*** AI code generation; https://www.sonarsource.com/learn/ai-code-generation/; last
access: May 04, 2025
*** AI Code Generation: The Risks and Benefits of AI in Software;
https://www.legitsecurity.com/aspm-knowledge-base/ai-code-generation-benefits-and-
risks; last access: May 04, 2025
*** AI Transparency: Fundamental pillar for ethical and safe AI; April 08, 2025;
https://www.plainconcepts.com/ai-transparency/; last access: May 04, 2025
*** Artificial Intelligence; January 30, 2025;
https://libguides.washburn.edu/c.php?g=1377087&p=10201459/; last access: May 04,
2025
*** Artificial intelligence and copyright;
https://en.wikipedia.org/wiki/Artificial_intelligence_and_copyright; last access: May 04,
2025
*** Artificial Intelligence and the Creative Double Bind - Chapter 2; Harvard Law
Review; Vol. 138, Issue 6, April 2025; https://harvardlawreview.org/print/vol-
138/artificial-intelligence-and-the-creative-double-bind/
*** Copyright Office Releases Part 2 of Artificial Intelligence Report;
https://www.copyright.gov/newsnet/2025/1060.html; last access: May 04, 2025
*** Generative AI Tools for Software Development | Coding Temple;
https://www.codingtemple.com/blog/generative-ai-tools-for-software-development/; last
access: May 04, 2025
*** How Java Is Different From Other Programming Languages;
https://www.fita.in/how-java-is-different-from-other-programming-languages/; last
access: May 04, 2025
*** Java (programming language) – Wikipedia;
https://en.wikipedia.org/wiki/Java_(programming_language); last access: May 04, 2025
*** Java Features; March 24, 2025; https://www.geeksforgeeks.org/java-features/; last
access: May 04, 2025

Pag. 327 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

*** The Future of Java Development: 13 AI Tools Transforming Productivity; January 15;
2025; https://groovetechnology.com/blog/technologies/ai-for-java-developers/; last
access: May 04, 2025
*** The Impact of Generative AI on Java Development: From Code to Deployment;
November 11, 2024; https://medium.com/@speckyfox.technologies/the-impact-of-
generative-ai-on-java-development-from-code-to-deployment-8d1e3d2a35d8; last access:
May 04, 2025
*** The Power of AI-Assisted Coding: Streamlining Rapid Development in Java;
September 11, 2024; https://blog.kodezi.com/the-power-of-ai-assisted-coding-
streamlining-rapid-development-in-java/; last access: May 04, 2025
*** What is AI code generation?; https://aws.amazon.com/what-is/ai-coding/; last access:
May 04, 2025

Pag. 328 / 350

Article’s total number of pages: 23
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

FINANCIAL ACCOUNTING INFORMATION IN THE CONTEXT OF SCORE-

BASED SUSTAINABILITY REPORTING ESG (ENVIRONMENTAL, SOCIAL,
GOVERNANCE)

Marilena ZUCA1
Alice Emilia ȚÎNȚA2
Anda Mihaela LĂCEANU3
Georgiana POPA4

Abstract
Reliability, accuracy, relevance, and compliance with current accounting regulations are all
necessary for reporting financial and accounting data responsibly from a sustainability
standpoint.
Sustainability refers to an organization's ability to meet long-term goals, which include
economic, social, and environmental aspects (ESG). Sustainable accounting is becoming
increasingly significant in today's world, as businesses are becoming more aware of their
environmental and societal impact and are more actively participating in sustainable
development initiatives. Reflecting accurate and relevant financial and accounting
information entails having a clear understanding of an organization's financial condition,
performance, and social and environmental impacts. Enterprises that have invested more in
social and environmental initiatives and have considered corporate governance have
reported increased profits. ESG activities benefit businesses at every stage of the value
chain, from cost reduction to maintaining a competitive advantage.
Keywords: sustainability, sustainability reporting, corporate social responsibility (CSR),
ESG scores
JEL Classification: M41

1. Introduction
Sustainable development is a big scientific field with untapped potential. Many academic
disciplines can contribute to the study of this topic, and accounting processes and actors in
the accounting business connect with many of these disciplines. Reality shows the
importance of analysing the function and impact of accounting methods within the

1PhD, Assoc. Professor, Romanian-American University, Romania, marilena_zuca@yahoo.ro, corresponding

author
2PhD, Lecturer, external public auditor, The Court of Accounts, Romania, alicetinta@yahoo.com
3PhD, director, external public auditor, The Court of Accounts, Romania, andastancioiu@yahoo.com
4PhD, external public auditor, The Court of Accounts, Romania, georgianapopa@yahoo.ro

Pag. 329 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

framework of sustainable development, using an interdisciplinary approach. This approach

may help us better understand how accounting might help promote sustainable development
and address environmental, public health, and poverty challenges. The application of the
idea of sustainable development necessitates adjustments in accounting practice to fulfil the
needs of users of financial and accounting data [1]. Although sustainable development is
becoming more prevalent in accounting and management literature, it is critical to define
this idea in the many contexts in which it is used [2]. According to research, the specialized
literature lacks empirical evidence supporting the existence of a direct causal link between
accounting and economic progress [3]. The International Federation of Accountants (IFAC)
emphasizes accounting and the accounting profession's critical role in achieving at least 8
of the 17 United Nations goals, such as improving educational quality, promoting gender
equality, improving working conditions, stimulating economic growth, supporting
innovation and infrastructure development, and promoting responsible resource
consumption.

2. Literature review
CSR (corporate social responsibility) reporting has been successively referred to as
'sustainability reporting', 'CSR reporting' (notably by the Global Reporting Initiative),
'extra-financial reporting' (the current official name), 'non-financial reporting' (the 2017
European name), 'extra-financial or non-financial disclosure', 'social or societal reporting',
etc. Behind this diversity nowadays is all the information that companies publish on the
social, societal and environmental aspects of their activities and their performance [4]. But
there may be other components in the future, as the gap will be widening between financial
information and non-financial. There is nothing to prevent, for example, the New Economic
Regulations (NER) information mentioned above. It should be noted that the 'governance'
aspect is sometimes included in extra-financial reporting and sometimes not because it has
been considered that it can be linked to accounting and financial information.

The regulatory process for sustainability reporting has evolved significantly in recent years.
According to the Financial Times, there are more than 230 initiatives on corporate
sustainability standards in more than 80 business sectors. According to a study [4] by Albu
N, the most widely used sustainability reporting frameworks are the Global Reporting
Initiatives (GRI) and the United Nations Global Compact (UNGC). GRI is mainly preferred
by multinational companies, while UNGC is preferred by small and medium-sized
enterprises due to its simpler nature. According to International Financial Reporting
Standards (IFRS), information is considered relevant when it can be used to make
predictions or estimates about an entity's future financial position and performance. In this
context, predictability is essential and relevant financial and accounting information can be
used to support decision making. Directive (EU) 2022/2464 is based on the premise that
information is performative, with the ability to create reality and influence behavior. This
distinguishes between constative and performative statements, requiring large, small and
medium-sized enterprises to include in their management reports information on their
sustainability impact and how it affects their development, performance and position. By
adopting this policy, companies make a commitment and promise, creating an obligation.
However, performance cannot be realized without a collective commitment between the
authors and receivers of the statements. Through its meaning, value, content and status,
Pag. 330 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Directive (EU) 2022/2464 encourages sustainability information to become more

performative, generating significant changes in the representation of actors and individuals.

3. Research methodology
The research methodology was based on a careful analysis of the function and influence of
accounting procedures in sustainable development, using an interdisciplinary approach.
This investigation sought to gain a better understanding of how accounting can help to
promote sustainable development and address difficulties linked to conveying business
performance in the social, environmental, and governance domains.
To obtain solid and complete conclusions, a hybrid strategy was used, integrating
quantitative and qualitative studies. To acquire a thorough knowledge of the complex
relationship between sustainability reporting and financial success in a set of companies in
the Reuters Eikon database, statistical data were supplemented with detailed field
observations. This combined research strategy enabled the identification of significant
elements impacting financial performance while also providing a comprehensive view of
the impact of non-financial information on the activities of the organizations under
consideration.
By evaluating the environmental impact and sustainability of enterprises (the application of
ESG standards) in the context of resource optimization, performance enhancement, and
decision-making, the focus was on quantifying sustainable value and sustainable added
value.
We list the following methods and techniques as part of the study's development: 1.
Documentary analysis and systematic review of specialized literature, as evidenced by
bibliographic and bibliographic references: theoretical documentation (books, national and
international articles, studies), as well as documentation pertaining to the applied part,
processing, and interpretation of information gleaned from the documentation;
Classification, synthesis, factor analysis, variation and correlation analysis, trend analysis,
average rate of variation, analysis of central tendency, average representativeness, and
graphic representation of the events and phenomena under investigation are examples of
mathematical and statistical techniques; 3. Interdisciplinary research approach: the study
draws on expertise from computer science, statistics, and finance and accounting. The
inquiry entails collecting, systematizing, analyzing, and interpreting data, studying specific
financial reports and publications, and formulating conclusions and suggestions.

4. Sustainability reporting - premises of enterprise performance

Sustainability reporting communicates a company's performance in the social,
environmental, and governance domains. This technique has been increasingly crucial for
businesses in recent years, as worries about their environmental and social impact have
grown. Sustainability reporting is regarded as a must-have for businesses seeking to
demonstrate their commitment to social responsibility while also attracting investors,
partners, and customers interested in sustainability (Table 1).

Table 1. Sustainability reporting requirements in relation to entity performance

Pag. 331 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Requirements Description

Sustainability reporting promotes transparency in business activities. Publishing

Transparency and
information about environmental impact and social contribution helps build trust
responsibility
among stakeholders, including customers, investors and employees.

By assessing and reporting sustainable performance, organizations can identify

environmental and social risks that could affect their business operations. This
Risk management
enables them to develop strategies to manage these risks and minimize potential
damage.
Adopting sustainability practices can stimulate business innovation. For example,
Innovation and
the development of environmentally friendly products and efficient production
competitiveness
processes can provide a competitive advantage in the marketplace.
By reporting performance in this field, businesses can more easily attract
Access to financing
investment and benefit from more advantageous financing conditions.
Organizations that have well-defined and effectively communicated sustainability
Employee involvement practices can attract and retain talented employees. An organizational culture
based on sustainable values can raise team motivation and engagement.
Many countries are adopting stringent environmental and social reporting
Regulations and
regulations. Companies that comply not only avoid penalties but also improve
compliance
their business reputation.
Sustainability reporting can help businesses contribute to global goals such as
Contribution to global
those of the United Nations 2030 Agenda for Sustainable Development. It enables
objectives
them to position their activities in the context of global issues.
Source: authors conception and processing

4.1. The influence of corporate social responsibility on the community and the
environment
The concept of corporate social responsibility (CSR) is a very modern conception, being
almost in the "embryonic stage”.[6] The definitions offered by different authors differ
significantly: some focus solely on the social dimension, while others highlight the
philanthropic and ethical aspects of commercial activity (particularly among American
authors). Some European authors emphasize the interconnectedness of the three pillars of
sustainable development: economics, society, and the environment.[7] Modern businesses
are encouraged to contribute to the common good by the growth of consumer and
environmental movements as well as rising interest group pressure. As a result, corporate
social responsibility (CSR) has emerged as a strategic problem for managers and a research
area for marketing and management science scholars in recent years. Peloza and Shang
claim that because of its influence on employee and customer behaviour, corporate social
responsibility, or CSR, is a subject of increasing managerial and scholarly attention.[8] It
also includes an opportunity to consider the company's relationship with society in a
broader context, as well as an environmental component. It can be defined as the managerial
process by which the company's strategic orientations toward sustainable development are
expressed. Under institutional and social pressure, most multinational corporations have
now launched a variety of social responsibility initiatives, originally with the goal of
addressing an image problem and enhancing commercial performance. For some
Pag. 332 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

businesses, taking social responsibility into account becomes a means of distinguishing

themselves from the competition; as a result, their innovation is focused on the development
of new goods and approaches that are more environmentally friendly. Investing in corporate
social responsibility becomes vital for businesses since it becomes a competitive aspect that
affects performance. The interaction with the new information and reporting requirements
in the field of corporate social responsibility (CSR) defines the need for significant
improvement/revision, or possibly a complete overhaul of the current accounting model.
This entails developing an accounting model tailored to the concept of "green capitalism".
Several recommendations have been made in this area, and extra-financial reporting is
becoming more regulated by legislative frameworks. According to Postel et al., CSR
appears to be an attempt to balance economic efficiency with ethical considerations. The
European Commission defines corporate social responsibility as the voluntary
incorporation of social and environmental issues into a company's commercial activities
and interactions with stakeholders. It is critical to consider the voluntary nature of this
behaviour. A lot of businesses aggressively highlight the ethical components of their
business practices. It is important to remember that many businesses are frequently under
pressure from stakeholders to adhere to these standards, even though responsible
commitments can be made voluntarily. The three essential elements of extra-financial
analysis are ESG criteria, which relate to the environment (such as conservation of
biodiversity), society (such as working conditions), and governance (such as business
ethics). These criteria enable the evaluation of how businesses incorporate the concepts of
sustainable development. For instance, AccountAbility bureau and its founder, J.
Elkington, spearheaded the adoption of the Triple Bottom Line strategy by numerous
multinational firms in the late 1990s. In order to supplement the evaluation of economic
success with an estimate of social consequences and an evaluation of how well the
company's operations correspond with the preservation of the planet's ecosystems, this
expression refers to the net result from the traditional profit and loss statement. As a result,
a metric that takes into account the enterprise's social and environmental effects in addition
to its financial performance is suggested. The "triple P"—which stands for "Profit, People,
Planet"—is another name for the approach. [9] In 1996, a consortium of multinational
corporations launched the Global Reporting organization (GRI), the oldest commercial
organization devoted to harmonizing social and environmental reporting, which utilized the
tertiary method. The social and environmental externalities—both positive and negative—
caused by economic activity are addressed by the GRI reference frameworks, which are
increasingly acknowledged as essential guidelines for non-financial reporting. The purpose
of the suggested non-financial indicators is to help with a better understanding and
management of enterprises. This does not mean arguing against the conventional
accounting model; rather, it means putting up an alternative framework that incorporates
non-financial data. The World Council for Sustainable Development's (WBCSD) affiliated
organizations seek to emphasize the private sector's pivotal role in promoting sustainable
development, a notion that has its origins in the political and macroeconomic domains.
The primary objective of these companies' experiments with different valuation methods is
to establish a price for natural capital so that externalities may be minimized, and it can be
managed. Therefore, expressing the value of natural capital is crucial to current discussions
on the evolution of social-environmental accounting models.

Pag. 333 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

4.2. Integration of non-financial reporting in Europe and Romania

As early as 2001, when Article 116 of the law on New Economic Regulations (NRE) was
implemented in France, and in 2014, when Directive 2014/95/EU on non-financial
reporting (also known as the Non-Financial Reporting Directive, or NFRD) was
implemented in Europe, a legal framework started to take shape, and the dissemination of
social and environmental information became mandatory. This directive was implemented
into French law by the decree of August 9, 2017, which requires major firms, both listed
and unlisted, to prepare an Annual Non-Financial Performance Statement (DPEF) and
include it in their annual management report. The information they must provide in this
statement relates to the measures they have taken to adapt to the consequences of climate
change, the voluntary objectives they have set themselves in the medium and long term to
reduce greenhouse gas emissions and the means they have implemented to this end, as well
as their actions aimed at combating discrimination and promoting diversity; listed
companies must also provide information on the protection of human rights and the fight
against corruption and tax evasion. However, this DPEF requirement is relatively vague
and lacks a formal framework; in essence, it does not link financial and non-financial
elements.
Non-financial reporting has grown in popularity among Romanian businesses as European
legislation evolves and investors and consumers seek greater openness. This type of
reporting involves the disclosure of information about a company's environmental, social,
and governance (ESG) performance, which supplements standard financial statistics and
provides a more comprehensive picture of sustainability and accountability. The
implementation of non-financial reporting has been considerably influenced by the
European Directive 2014/95/EU, which requires certain types of large enterprises to publish
non-financial information.
According to current legislation (OMFP no. 85/2024, issued in the Official Gazette no. 75
on January 26, 2024), enterprises with more than 250 workers, a net turnover of more than
50 million euros, or total assets of more than 43 million euros must submit these reports.
Some of the important features of non-financial reporting in Romania are:
➔ Transparency and responsibility: Non-financial reporting contributes to increasing
transparency within companies, providing stakeholders with relevant information about
their impact on the environment and society.
➔ Research and innovation: By reporting on needs and actions taken in ESG areas,
businesses can identify areas that require improvement, thus being encouraged to invest in
research and innovation to develop more sustainable practices.
➔ Attracting investment: More and more investors are looking for investment
opportunities in companies that demonstrate their commitment to sustainability. Non-
financial reporting provides an assessment of the risks and opportunities associated with a
company's business practices.
Although the importance of non-financial reporting is clear, Romanian companies face
various challenges in implementing this process. These include:
➔ Lack of Unified Standards: Variability in reporting standards and frameworks can
create confusion and difficulties in assessing information. Businesses must decide which
standards to adopt, and the absence of a clear national framework can lead to
inconsistencies.
Pag. 334 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

➔ Administrative capacity: Many small and medium-sized enterprises lack the

resources to collect, analyse and publish non-financial information in a coherent and
efficient manner.
➔ Organizational culture: The implementation of non-financial reporting necessitates
a shift of thinking inside companies, which can be a challenge, particularly in traditional
businesses or those that have not previously been exposed to transparency standards.
➔ Public perception: There is little consumer awareness or interest in non-financial
data. This might make it challenging to persuade businesses to engage in non-financial
reporting because the benefits are not always evident or understood.
➔ Limited resources: Many businesses may not have the financial or human resources
to create a successful non-financial reporting system. They may believe that the costs
connected with creating these reports outweigh the purported benefits.
➔ Lack of expertise: Companies may lack knowledge and skills in the area of non-
financial reporting. Training employees and developing proper internal processes for
gathering and analysing non-financial data can be difficult.
➔ transparency: Some companies are concerned that having a clear position on non-
financial reporting may put them at a competitive disadvantage. This can lead to reluctance
to implement such a method, which is viewed as risky.

4.3. The role of non-financial information in presenting the performance of the

enterprise
The 2008 financial crisis exposed the risks associated with an overemphasis on short-term
financial performance, emphasizing the necessity for a different perspective on entity
performance. This should encompass financial stability, long-term success, and
sustainability. Numerous worldwide and national initiatives emphasize the relevance of
large firms' involvement in society, in addition to the traditional goal of generating
shareholder income. [10] One pertinent example is the United Nations Global Compact,
which was established in 2000 and supports 10 globally agreed principles in the areas of
human rights, labour, the environment, and anti-corruption. Another notable endeavour is
the ISO 26000 standard, which was adopted in 2010. It provides recommendations on how
corporations can be socially responsible. Furthermore, national initiatives, such as the
German Sustainability Code, established in 2011, coexist with international standards,
demonstrating a global trend toward corporate social responsibility. These projects
demonstrate the necessity for management to give information about the enterprise's social,
environmental, and intangible activities and performance.
In addition, they clearly address the information needs of stakeholders other than
shareholders. While financial information is subject to well-established accounting
standards such as Generally Accepted Accounting Principles and other securities-specific
laws, non-financial information (NFI) is less governed by a consistent regulatory
framework. However, in recent decades, a number of projects have emerged that offer
guidance on how to effectively communicate NFI in financial markets. One such response
comes from the Global Reporting Initiative, a non-profit organization in the United States
founded in 1997 to provide guidelines for sustainability reporting, which is defined as a
report published by a company or organization on the economic, environmental, and social
impacts of its activities. Another example is the International Integrated Reporting Council
(IIRC), which was established in 2010 with the goal of encouraging companies to transform
Pag. 335 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

their reporting by incorporating various information on value generation. In 2013, the IIRC
published the Integrated Reporting Framework, which makes it easier to create integrated
reports by providing a concise presentation of how an entity's strategy, governance,
performance, and prospects contribute to value creation in the short, medium, and long
term, within the context of its operating environment.
In 2013, the European Commission responded to the need for non-financial reporting
(NFR) by amending accounting legislation to require large corporations to boost
transparency about their social and environmental performance. As investors and other
stakeholders grow more conscious of the social and environmental consequences of
economic operations, non-financial data becomes increasingly important in providing a
complete picture of a company's success.
In this light, a substantial shift has occurred in the way investors evaluate companies.
Instead of concentrating just on short-term financial metrics, they are shifting their focus to
larger factors that can affect long-term performance. As a result, investment evaluation and
selection are becoming more sophisticated processes that incorporate environmental, social,
and governance (ESG) variables into traditional analysis, contributing to a more holistic
view of firm performance. As a result, the concept of firm performance is broadening
beyond traditional financial indicators, incorporating non-financial characteristics that
reflect a more comprehensive and responsible vision of long-term success.
This development emphasizes the importance of investors and corporate managers
rethinking their performance perspectives and adopting a comprehensive approach that
balances financial interest, social responsibility, and sustainability.

4.4. Representation of financial and accounting data in environmental accounting

Environmental accounting is a technique that allows organizations to quantify the outcomes
of their efforts to transform the economic system in accordance with the principles of
sustainable development [11]. Environment accounting also known as green accounting
[12], is an extension of the System of National Accounts that includes the usage and
depletion of natural resources. This type of accounting is critical for controlling the
environmental and operational costs of natural resources. Natural resource valuation is
important in analysing social costs and benefits, as well as in certain parts of environmental
accounting. Environmental accounting has several categories, each concentrating on a
different aspect of an entity's contact with the environment. There are several major types:
● Accounting of physical or natural resources
This type entails accounting for the quantity and quality of natural resources used by a
company. Physical resource accounting aids to sustainable development planning by
including environmental issues into economic studies, thereby balancing economic
growth and environmental conservation. It aids in tracking the depletion or enhancement
of natural resources, providing data on their availability and sustainability.
● Environmental costs accounting
Accounting is the process of identifying, measuring, and analyzing the expenses connected
with a company's environmental actions. It comprises both internal and external costs
associated with environmental issues such as pollution, resource depletion, and waste
management. It enables firms to assess the financial implications of their environmental
activities, such as the costs of pollution control, waste management, and compliance.
Pag. 336 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

● Analysing eco-efficiency
It evaluates the efficacy of resource utilization in manufacturing processes and identifies
possibilities to improve environmental performance. This methodology seeks to create
sustainable development by optimizing resource utilization, reducing environmental
impact, and enhancing economic value. It assists organizations in optimizing their
operations to reduce resource consumption, waste generation, and environmental impact.
• Greenhouse gas accounting
It is the sort of accounting that involves monitoring and reporting greenhouse gas (GHG)
emissions from a company's operations. GHG accounting is critical for an organization's
sustainability, regulatory compliance, and worldwide efforts to prevent climate change. It
tackles climate change concerns by enabling organizations to measure and manage their
carbon footprint.
• Environmental performance indicators
They are quantitative measures for evaluating and communicating a company's
environmental performance. These indicators give vital information about environmental
impacts and sustainability practices, allowing stakeholders to evaluate progress, set goals,
and make educated decisions. They include energy usage, water use, trash generation, and
emissions, which provide an overview of environmental effect.
● Reporting stable development
The disclosure of an organization's social, environmental, and economic performance is
known as sustainability reporting. In order to better understand how the organization
incorporates sustainability into its operations and decision-making, pertinent information,
statistics, and initiatives must be disclosed as part of this reporting process. It offers a
comprehensive perspective on how a business operates, accounting for the triple bottom
line—economic, environmental, and social factors.
● Full cost accounting
It considers both direct and indirect expenses connected with an entity's operations,
including environmental and social costs. Full Cost Accounting aims to provide a thorough
awareness of the true economic and environmental costs associated with decision-making.
It seeks to provide a more thorough knowledge of production costs, including externalities
that may affect the environment.
● Biodiversity accounting
It entails evaluating how an entity's operations affect ecosystems and biodiversity. This
accounting approach measures and evaluates the diversity of species, their relationships,
and their ecological responsibilities in a way that goes beyond conventional financial
reporting. It aids organizations in comprehending and controlling their roles in biodiversity
conservation or loss.
● Corporate environmental reporting
It entails sharing environmental performance statistics and related information with
stakeholders. It necessitates open communication of a company's environmental actions,
practices, and accomplishments to all stakeholders, including investors, customers,
employees, regulators, and the general public. It promotes transparency and accountability,
Pag. 337 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

allowing stakeholders to make informed judgments about a company's environmental

practices.
Romania, as a member of the European Union, has begun to place a greater emphasis on
sustainability reporting, considering the considerable impact of economic operations on the
environment and society. In light of climate change and more stringent international
regulations, Romanian businesses and organizations are encouraged to develop open
sustainability strategies. As a result, a number of regulations aimed at sustainability
reporting have been implemented at the national and European levels, including the
European Union Non-Financial Reporting Directive (NFRD), which requires certain
entities to disclose information about their environmental, human rights, and corruption
impacts. This law was updated in 2021 and renamed the Corporate Sustainability Reporting
law (CSRD), which expanded reporting requirements even further. Many Romanian
companies are already structuring their sustainability reports using international standards
such as GRI (Global Reporting Initiative) or SASB (Sustainability Accounting Standards
Board). These guidelines provide a systematic framework for organizations to convey their
economic, social, and environmental impacts. However, sustainability reporting in
Romania faces a number of obstacles. Some small and medium-sized businesses lack the
resources and skills required to create an efficient reporting system, and they have a limited
understanding of the idea of sustainability.

4.5. ESG effects on performance in terms of added value

The notion of ESG (Environmental, Social, Governance) is becoming increasingly
important in evaluating the performance of companies. Its emergence is attributed to a
growing awareness of the impact of economic operations on the environment, society, and
corporate governance. Implementing ESG standards improves organizations' financial and
non-financial performance, demonstrating the impact of sustainable values on economic
outcomes. Sustainable value is a notion that considers economic, social, and environmental
outcomes when evaluating a company's long-term success.
It is described as an entity's potential to earn money while also positively contributing to
society and the environment. ESG standards are used to assess performance in these three
areas and serve as an analytical framework for investors and management. Sustainable
value added is defined as the additional value gained while keeping the total level of
environmental and social consequences constant [13]. According to studies, there is a good
association between ESG (Environmental, Social, and Governance) practices and financial
performance of entities. Friede's research, which evaluated over 2000 empirical studies,
found that 63% of them indicated a positive relationship between ESG and financial
performance. Financial information created by businesses that apply effective ESG
initiatives, such as enhancing energy efficiency or cooperating with local communities, can
help cut operational costs and boost consumer loyalty, according to the McKinsey analysis
[14]. Companies with strong ESG scores experienced 3%-5% better sales growth than
those with poor scores. Entities that incorporate environmental, social, and governance
(ESG) concepts into their business strategy can dramatically minimize operational and
reputational risk. Effective governance management may prevent concerns like
information leakage, corruption, and fraud, while implementing sound environmental
policies can lower the chance of future penalties or litigation. According to a Harvard
Business School report published in 2021 [15], organizations that demonstrate appropriate
Pag. 338 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

ESG risk management are better prepared to deal with economic catastrophes, such as the
COVID-19 outbreak.

5. Discussion and findings of a study on the use of financial and accounting

information in generating ESG scores
An effective and quick evaluation of the non-financial performance of businesses has been
made possible by access to a variety of databases, including ASSET4, EIKON,
Sustainalytics, MSCI ESG (KLD), and Bloomberg. These databases also provide the ability
to compare organizations, sectors, or nations. As a result, it is anticipated that the need for
ESG data will only increase, and these databases will aid investors in making investment
decisions. The number of entities whose data will be available from these ESG databases
will surely increase as a result of the European Union's legislative amendments pertaining
to the disclosure of non-financial information. The Thomson Reuters EIKON database
gives access to trustworthy, current, and accurate data from over 400 capital markets and
over-the-counter markets. The database contains a wide range of financial indicators and
information (both current and archived), including stocks and bonds, investment and trust
funds, exchange rates, interest rates, financial derivatives, and commodities (raw materials),
as well as international macroeconomic data and forecasts for the world's largest economies
and developing countries. The Thomson Reuters EIKON database contains:
• ESG data and results for over 6000 companies,
• over 400 partial data reported during the sustainable development,
• over 70 KPIs (key performance indicators),
• data from the year 2002,
• ESG data collected in real time from 75 000 sources,
• solutions that allow monitoring and reporting CO2 emissions to fulfil legal requirements.
In this context, the Thomson Reuters EIKON database allows you to access ESG data on a
company and, eventually, an ESG rating for that business.
The calculation of the indicator value that classifies a company into an appropriate score is
based on three factors:
𝑎+𝑏/2
➔ ➔ indicator for the score ESG =
𝑐
where: a - the number of underperforming companies,
b - the number of companies with the same results as the one evaluated,
c - total number of companies with results.
Based on the calculated result, the company is assigned a specific ESG score on a scale
from D- to A+.
There are three categories of ESG indicators in the Thomson Reuters EIKON database:
- ESG Score,
- ESG Controversy Score (ESGC Score),
- ESG Combined Score.
The ESG score evaluates organizations' ESG performance using publicly available data
from 10 subject categories. Thomson Reuters compiles and analyses over 400 ESG data
points about a company, selecting 178 essential data elements for the final ESG score. The
data obtained is based on criteria such as materiality, data availability, and sector
significance. The data analysis allows us to state that the following categories can be
Pag. 339 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

included in the group of five important categories (taking the relevance of each category as
a criterion) that have the biggest impact on the company's ESG score:
• management (importance– 19%) – pillar: governance,
• labour (importance – 16%) – pillar: social,
• emissions (importance– 12%) – pillar: environment,
• use of resources (importance – 11%) – pillar: environment,
• innovation (importance– 11%) – pillar: environment.
The five categories stated above have a total relevance of 69%. It is also worth noting that
the major categories contain all of the environmental pillar's areas. It is also worth noting
that a company's CSR strategy (together with human rights) is the least important factor in
the ESG score, accounting for only 4.5%. The ESG Controversy Score (ESGC Score)
assesses a company's exposure to environmental, social, and governance issues, as well as
negative news coverage in the worldwide media. The Reuters Eikon database includes
3107 firms from advanced and emerging nations that operate in a variety of industry
categories (table 2).
Table 2. Descriptive statistics
Mean Median Stdev Min Max
ESG_score 51,99 53,58 20,40 0,47 95,06
ESG_Combined_Score 49,89 50,85 19,38 0,47 94,59
Environmental_Pillar_
Score 47,70 50,32 27,79 0,00 99,18
Social_Pillar_Score 52,55 53,88 23,99 0,26 98,63
Governance_Pillar_Sco
re 54,87 56,96 22,04 0,29 99,51
Size (ln_assets) 22,13 22,42 2,84 0,00 28,70
ROA 4,60% 3,34% 7,44% -113,99% 249,32%
ROE 12,17% 10,07% 72,30% -5332,55% 2604,97%
1460694,
Debt_to_Equity 209,69% 58,15% 10043,60% -0,55% 54%
Source: own editing

The table depicts the ESG (Environmental, Social, and Governance) performance, size, and
financial performance of a sample businesses from the Reuters Eikon database. The
calculations yield an average ESG score of 51.99, indicating moderate ESG performance
overall. The median, which is somewhat higher than the average, is 53.58, indicating a
slightly better-scoring dispersion. ESG scores vary significantly, with a standard deviation
of 20.40, indicating a large disparity between companies. The extreme numbers, 0.47 to
95.06, demonstrate significant disparities in ESG performance, ranging from very poor to
very high. The combined ESG score averages 49.89, which is slightly lower than the
average of the individual ESG scores. This could indicate a tiny discrepancy in how the
performances are aggregated or reported. The median total ESG score of 50.85, which is
close to the average, indicates a fairly balanced distribution.
Pag. 340 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The environmental pillar scores have a mean of 47.70 and a median of 50.32, indicating a
slightly centred upper end with a larger variation (standard deviation of 27.79). The mean
for the social pillar is 52.55 and the median is 53.88, indicating that social performance is
stronger on average than environmental performance, but there is also significant
variability, as evidenced by the standard deviation of 23.99. The governance pillar has the
best relative performance, with a mean of 54.87 and a median of 56.96, indicating that most
entities effectively handle these components, albeit significant variation.
The average size of the entities, as measured by the logarithm of assets (ln_assets), is 22.13,
with a median of 22.42 and a standard deviation of 2.84. This suggests a more consistent
size for these entities. Financial performance, as measured by return on assets (ROA) and
return on equity (ROE), demonstrates a diverse position. The ROA has a mean of 4.60%
and a median of 3.34%, indicating modest financial performance for most organisations.
However, the standard deviation of 7.44% and extreme values (minimum -113.99% and
maximum 249.32%) show significant variations amongst entities. The average return on
equity (ROE) is 12.17%, with a median of 10.07%. However, there is significant variance
with a standard deviation of 72.30% and extreme values ranging from -5332.55% to
2604.97%.
The debt-to-equity ratio varies the most, with an average of 209.69% and a median of
58.15%. The large standard deviation of 10043.60%, combined with extreme values
(ranging from -0.55% to 1460694.54%), indicates a significant difference in capital
structure between firms. This could indicate severe financial risk for some or strong
capitalisation for others. This research captures the complexity and diversity of the entities'
ESG and financial performance.
By analysing ESG ratings in relation to ROA and ROE, an increasing trend can be noticed,
indicating that companies with sustainable and socially and environmentally responsible
policies perform better financially. However, the substantial variability in both data sets
implies that this is not a general rule, and that each company has unique factors that
influence both ESG score and financial viability. It is critical that investors and stakeholders
continue to analyse these relationships in order to make sound decisions. Table 3 shows the
evolution of the average ESG score for the analysed organisations from 2018 to 2024.
Table 3. ESG score values, 2018-2024

Source: own editing

The overall ESG score has steadily increased from 45.05 in 2018 to 58.48 in 2024. This
trend shows that corporations' environmental, social, and governance performance is
Pag. 341 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

continuing to improve, potentially due to increased pressure from investors, regulators, or

public awareness of the relevance of these issues. The combined ESG score follows a
similar trend, but its values are significantly lower than the overall ESG score. This may
indicate that the combined company-level performance is slightly poorer than the separate
ones on different dimensions.
Looking more closely at the ESG pillars, it is clear that all three components
(environmental, social, and governance scores) have improved over the years. The
environmental pillar, for example, has risen from 41.26 in 2018 to 54.63 in 2024, indicating
a more widespread adoption of sustainable methods in business operations. Similarly, the
social and governance pillars have grown significantly, reflecting an increased interest in
social responsibility and corporate governance excellence.
The company size, as assessed by the logarithm of assets (ln_assets), remained largely
consistent during the analysed period, with only minor changes between 21.99 and 22.27.
This constancy shows that firm size did not play a significant role in the variances in ESG
rankings and financial performance.
ROA (return on assets) and ROE (return on equity) indicate little volatility, with ROA
values ranging from 3.82% to 5.10%. This indicates a relatively consistent operational
efficiency. On the other hand, ROE changes more, from 8.45% in 2019 to 13.74% in 2014,
demonstrating greater variability in return on equity, which can be influenced by both
capital structure and net income swings.
The evolution of the Debt-to-Equity ratio, which demonstrates significant volatility, is one
of the analysis's highlights. Although the average of this indicator shows a high level of
indebtedness, substantial swings, such as the sudden increase to 603.25% in 2017, signal
periods of financial instability or significant changes in a company's capital structure. This
could indicate different financing tactics used by businesses in response to economic
opportunities or problems throughout the relevant time. Overall, the data given show an
improvement in ESG performance over time, which is associated with a rather moderate
variation in financial success. The rise in ESG scores indicates that companies are becoming
more committed to sustainable and responsible practices, though the stability of company
size and the volatility of financial indicators suggest that these improvements are not
necessarily the result of company growth or stable financial performance, but rather of
increased awareness and external pressure for better corporate governance and social
responsibility (graph 1).
Graph 1. ESG score values, 2018-2024

Source: authors editing

Pag. 342 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

The graph shows the evolution of the ESG score from 2018 to 2024. A definite rising trend
can be seen, indicating a consistent improvement in organisations' ESG performance over
the period studied. Starting in 2018, the ESG score was 45.05, but by 2024 it had risen
dramatically to 58.48. This constant increase reflects firms' growing commitment to
incorporating environmental, social, and governance policies into their daily operations.
The fact that each year sees an increase over the previous year indicates a stable and
persistent tendency, rather than a transitory reaction to specific external or internal
circumstances. This trend could be attributed to a variety of factors, including stronger
regulation of ESG reporting and performance, a rise in investor demand for openness and
accountability, or firms' improved understanding of their influence on society and the
environment.
The rise can also be ascribed to the continual improvement of corporate governance, as well
as the policies and procedures put in place to promote sustainability. The data shows a
positive and significant evolution of ESG scores, demonstrating a gradual but consistent
shift by corporations towards more sustainable and responsible practices over the period
studied. This evolution shows that ESG factors have grown more incorporated into
business planning, possibly becoming vital for long-term success.
The average profitability of businesses has evolved in the sense of those described in graph
2. We see a correlation between ROA and ROE in the sense that assets used in business
and capital invested are stable, with net profit influencing enterprise success.

Graph 2. ROA and ROE evolution, 2018-2024

Source: authors editing

6. Conclusions
Sustainable development has been defined as meeting current needs without jeopardising
future generations' ability to meet their own, with the goal of striking a balance between
economic growth, social inclusion, and environmental protection, and serving as the
foundation for all EU policies and initiatives. Under Article 3(3) of the Treaty on European
Union, the EU officially recognises sustainable development as a long-term goal.
Pag. 343 / 350
Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Sustainability reporting is more than simply a legal requirement; it is also a strategic

opportunity for businesses to improve their performance and provide value in the long term.
Businesses can help to create a more sustainable future by incorporating the principles of
sustainability reporting into their operating plans.
Companies that invested more in social, environmental, and corporate governance
initiatives saw higher returns. In summary, despite the additional costs associated with
these measures, their revenues climbed, leading to the conclusion that organisations
engaging in ESG actions will benefit in the short and medium term. ESG activities benefit
businesses at every stage of the value chain, from cost reduction to maintaining a
competitive advantage. ESG encourages investors and publicly traded companies to
consider sustainability, resulting in a more robust green stock market.

References
[1] Dumitrana, M., Jianu, I., Lapteş, R., Popa, A. F. (2009), Sustainable development and
environmental accounting: concepts, trends and quality of accounting information, Journal
of Accounting and Management Information Systems, vol. 8, nr. 1, pp. 27-39
[2] Godemann, J., Bebbington, J., Herzig, C., Moon, J. (2014), Higher education and
sustainable development. Exploring possibilities for organisational change, Accounting,
Auditing and Accountability Journal, vol. 27, nr. 2, pp. 218-233
[3] Venter, E.R., Gordon, E.A. şi Street, D.L. (2018), The role of accounting and the
accountancy profession in economic development: A research agenda, Journal of
International Financial Management and Accounting, vol. 29, nr. 2, pp. 195-218
[4] Igalens, J. (2023) , 4. La reddition de comptes en matière de RSE, in Les grands auteurs
francophones. EMS Editions, pp. 69–98.
[5] Albu, N., Albu, C.N., Dumitru, M., Dumitru, V.F. (2013), Plurality or convergence in
sustainability reporting standards, Amfiteatru Economic, vol. XV, nr. 7, pp. 513-527
[6] Armstrong, M. (2006), A Handbook of Human Resource Management Practice, 10th
Edition, Kogan
Page Publishing, London.
[7] Aßländer, M.S. (2011), Corporate Social Responsibility as Subsidiary Co-
Responsibility: A Macroeconomic Perspective. Journal of Business Ethics, 99 (1), 115-128.
[8] Peloza, J., Shang, J. How can corporate social responsibility activities create value for
stakeholders? A systematic review. J. of the Acad. Mark. Sci. 39, 117–135 (2011).
https://doi.org/10.1007/s11747-010-0213-6.
[9] Bernard Colasse, Frédérique Dejean. Représentation comptable de l’entreprise et
développement durable. L’Économie politique, 2022.
[10] Erkens, Michael, Luc Paugam, and Hervé Stolowy. Non-financial information: State
of the art and research perspectives based on a bibliometric study. Comptabilité Contrôle
Audit 21.3 (2015): 15-92.

Pag. 344 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

[11] Nicolas Antheaume, Bernard Christophe, La comptabilité environnementale, des outils

pour évaluer la performance écologique, E-theque, 2005.
[12] Iyyanki V. Muralikrishna, Valli Manickam, in Environmental Management, 2017.
[13] Figge, F., Hahn, T. (2004a), Sustainable Value Added: measuring corporate
contributions to sustainability beyond eco-efficiency, Ecological Economics, Vol. 48, No.
2, pp. 173–187, doi: 10.1016/j.ecolecon.2003.08.005
[14] McKinsey & Company. (2020). How ESG affects financial performance.
[15] Harvard Business School. (2021). The Impact of ESG on Business Resilience During
Crisis

Bibliography
-books and manuals
Brogi, M., Lagasio, V., (2019), Environmental, Social, and Governance and Company
Profitability: Are Financial Intermediaries Different?, Corporate Social Responsibility and
Environmental Management, Volume 26(3), pp. 576–587
Burlaud, A., Niculescu, M. (2015), Non-financial reporting: a European perspective, Audit
financiar XIII Nr. 6 (126), p. 102-112.
Déjean F., (2021), Responsabilité sociale de l’entreprise et performance intégrée : la course
aux indicateurs, In L’Etat du Management, coll. Repères, La Découverte, p. 38-48.
Martinet A.C. et E. Reynaud (2004), Stratégies d’entreprise et écologie, Economica, Paris.
Postel N., Rousseau S. et Sobel R. (2006), La « responsabilité sociale et environnementale
des entreprises» : une reconfiguration potentielle du rapport salarial fordiste ?, Économie
appliquée, 4, 77-104.
Ribando J.M., Bonne G., (2010), A new quality factor: Finding alpha with ASSET4 ESG
data, Thomson Reuter, Starmine Research
Trinquecoste J.F., (2008), Responsabilité, Ethique et Logique marchande, Collection
Questions de société, éditions EMS.
-legislation and studies
Directive (EU) 2022/2464 of the European Parliament and of the Council of December 14,
2022 amending Regulation (EU) No 537/2014, Directive 2004/109/EC, Directive
2006/43/EC and Directive 2013/34/EU as regards sustainability reporting by undertakings.
OMFP nr. 85/2024, Monitorul Oficial nr. 75 din 26 ianuarie 2024.
Wikipedia contributors (2024) ISO 26000, Wikipedia, The Free Encyclopedia. Available at:
https://en.wikipedia.org/w/index.php?title=ISO_26000&oldid=1242973341
System of monitoring based on methodically organised accounts, representing the size of
economically valuable and limited reserves of natural resources and using physical
quantifiers such us tonnes or cubic metres, Eurostat. 2020. Economy-wide material flow
accounts and derived indicators. Luxembourg.

Pag. 345 / 350

Article’s total number of pages: 18
 Journal of Information Systems & Operations Management, Vol. 19.1, May 2025

Emissions trading scheme: targeting the free allocation of allowances (no data) Europa.eu.
Available at: https://op.europa.eu/webpub/eca/special-reports/emissions-trading-system-
18-2020/ro/
Set of performance indicators and standards Ministry of Environment, Water and Forests,
available at https://www.mmediu.ro/app/webroot/uploads/files/2020-11-
13_Studiu_indicatori_standarde.pdf
Full cost accounting (2017) European Environment Agency. Available at:
https://www.eea.europa.eu/help/glossary/eea-glossary/full-cost-accounting
HG nr. 1117 /2023 on the approval of the Methodology for Sustainability Reporting.
Romanian Code of Sustainability issued by the Government of Romania Published in the
Official Gazette No. 1052 of November 21, 2020.
Rapid case review Reporting on sustainability: A stocktake of EU Institutions and Agencies
- Europa.eu.
Global Sustainable Investment Alliance (GSIA). (2021). Global Sustainable Investment
Review 2020.
Thomson Reuters, 2017c, Thomson Reuters ESG data and solutions, https://financial.
thomsonreuters.com/content/dam/openweb/documents/pdf/financial/esg-research-
brochure.pdf

Pag. 346 / 350

Article’s total number of pages: 18
JOURNAL OF INFORMATION SYSTEMS & OPERATIONS MANAGEMENT
VOL. 19 NO. 1 / 2025