Artificial Intelligence Principles

Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

Version History

Department:
Version Date Section(s) Description
V1.0 22/Mar/2024 All First version

Author Reviewer Approver

Michel Lemos Daniel Baird, Barry Johnston, Ian Lee Richard Abraham
and Loren Kelly
(The Reviewer must never be the author, but the author and reviewer can be the approver)

Revision History
Author (A) / Reviewer (R) / Approver (AP) Section(s) Comments (version, date, and brief details)
A: Michel Lemos
R: Daniel Baird, Barry Johnston, Loren Kelly and Ian
V1.0 [22/Mar/2024] – First version
Lee
AP: Richard Abraham
A:
R: vx.x [dd/Mmm/YYYY] –
AP:
A:
R: vx.x [dd/Mmm/YYYY] –
AP:
Note: Any Document in paper must be used for reference purposes only. The on-line copy must be considered the current document.

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 2
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

Scope
The primary objective of this document is to set out Petrolink Group’s position on
artificial intelligence (AI), with a view to implementing AI’s full potential whilst
mitigating the risks of its use. This document sets out guidelines for how to use and
develop AI technology.

This document also helps to produce informed decisions within the Petrolink Group
on the following issues: how to retain a human-centric approach to Petrolink Group’s
business, inclusivity, biases and fairness, data privacy, transparency, security and
compliance with relevant laws and regulations.

Background
The nature of AI is itself dynamic, and so customer expectations around AI systems
are high.

The Petrolink Group acknowledges that the use of AI can be a positive tool to many
organisations, and it has already been put to good use within the organisation.
However, there are some inherent risks when using AI systems. For example:

• The lack of traceability and interpretability in AI algorithms can hinder understanding

of how AI systems arrive at their decisions raising concerns about transparency,
accountability, and trustworthiness.
• AI-generated content of any nature may not be unique, and the direct copying of it
could result in copyright infringement.
• AI models may be susceptible to inputs deliberately designed to deceive or
manipulate AI systems, posing security risks and undermining their reliability in
safety-critical applications.

Whilst the above is not an exhaustive list of the risks present when using AI systems,
the key point is that use of AI systems can give rise to various legal, compliance and
commercial risks that the Petrolink Group wishes to avoid.

Thus, there is a need to implement applicable controls concerning use of AI - controls

covering areas such as security, safety, fairness, transparency, data quality and
quality of AI systems throughout their life cycle.

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 3
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

An AI system is a machine-based system that is capable of influencing the

environment by producing an output (predictions, recommendations, or decisions) for
a given set of objectives. AI systems are designed to operate with varying levels of
autonomy.” (OECD definition).
OECD.AI

AI is a technical and scientific field devoted to the engineered system that generates
outputs such as content, forecasts, recommendations or decisions for a given set of
human-defined objectives. of the inherent dignity and of the equal and inalienable
rights of all members of the human family is the foundation of freedom, justice, and
peace in the world.
ISO/IEC 22989:2022

The theory and development of computer systems able to perform tasks normally
requiring human intelligence, such as visual perception, speech recognition, decision-
making, and translation between languages. Modern AI is usually built using machine
learning algorithms. The algorithms find complex patterns in data which can be used
to form rules.
UK AI Safety Institute

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 4
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

Abbreviations
AI Artificial Intelligence
EU European Union
IEC International Electrotechnical Commission
ISO International Organization for Standardization
NIST National Institute of standards and Technology
OECD The Organization for Economic Cooperation and Development
UK United Kingdom

Definitions
Explainable AI - a set of processes and methods that allows human users to comprehend and
trust the results and output created by machine learning algorithms. Explainable AI is used to
describe an AI model, its expected impact and potential biases.
Petrolink Group / Petrolink - Petrolink International Limited, its subsidiaries, its holding
company or its affiliates. Collectively these organizations are known as the “Petrolink Group“ or
”Petrolink”.
Policy - intentions and direction of an organization as formally expressed by its Top
Management. (ISO/IEC 42001:2023)
Stakeholders – any individual, group, or organization that can affect, be affected by, or
perceive itself to be affected by a decision or activity. (ISO/IEC ISO 22989:2022)
Top Management - person or group of people who directs and controls an organization at the
highest level. (ISO/IEC 42001:2023)

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 5
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

Table of Contents
Scope ............................................................................................................................................... 3
Background ..................................................................................................................................... 3
Abbreviations .................................................................................................................................. 5
Definitions........................................................................................................................................ 5
1. Introduction .............................................................................................................................. 7
2. AI Principles.............................................................................................................................. 7
3. Statement Policy....................................................................................................................... 8
4. Expectations ............................................................................................................................. 8
5. Further reading ......................................................................................................................... 9
6. References ................................................................................................................................ 9

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 6
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

1. Introduction
The Petrolink Group, through the fundamental properties of the four principles of Explainable
AI (figure 1), is committed to:
• Advancing AI technology.
• Reducing biases towards AI technology.
• Collaborating with industry partners to mitigate potentially harmful uses of AI.

Petrolink Group fully supports trustworthy approaches based on international standards and
frameworks, such as ISO/IEC, NIST and other regulatory agencies which provide essential
guidance towards trust and safety.

Petrolink Group understands the main purpose of AI is to enhance human intelligence and not to
replace human intelligence. Furthermore, Petrolink understands the important role AI can play in
protecting the privacy of its clients and is committed to supporting initiatives regarding algorithm
transparency and appropriate design confidence in its outputs.

Figure 1 – Four principles of Explainable AI, NIST.IR.8312

These fundamental properties can include, but are not limited to, regulatory and legal
requirements, quality control of an AI system, and customer relations.

2. AI Principles
The Petrolink Group commits to ensuring the following principles are upheld:
• Transparency to reinforce trustworthiness.
• Promotion of inclusivity and fairness to individuals or groups of individuals, and
elimination of human biases.
• Prioritization and safeguarding of Stakeholders’ privacy and data rights.
• Active defence from adversarial attacks - thereby minimizing security risks and
improving confidence in system outcomes.

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 7
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

3. Statement Policy
Petrolink group’s position on AI is as follows - Within its business environment and systems,
Petrolink Group:
- Supports AI initiatives that safeguard the rule of law, democratic institutions, and
fundamental rights.
- Supports the use of AI founded on ethical behaviour, based on a risk-based approach to
conform with present laws and the continual deployment of regulations and standards.
All of these represent components of continual improvement, including awareness of the
AI dynamic ecosystem.
- Encourages the use of business AI to foster innovation, human capital and information
security while ensuring safety and protecting human rights.
- Fosters an environment where AI development thrives across cultural diversities.
- Ensures AI technologies cater to the unique needs and contexts of different societies.
- Works towards preventing sociotechnical AI risks of AI misuse including disinformation,
harmful actions, and societal harms, and reduce unfair biases.
- Uses AI to provide trustworthy information to Stakeholders based on statutory and
regulatory frameworks.
- Develops and deploys AI safely and responsibly, outlining security risks and prohibiting
the input or use of sensitive data within AI tools.
- Implements measures across the business to ensure AI tools are reviewed and
approved by Information security management.
- Prohibits utilizing or entering any:
o commercially sensitive or confidential business information; and
o personally identifiable information into AI systems.
- Provides training and support to employees to ensure the above objectives are met.

4. Expectations
The Petrolink Group expects that all Stakeholders :
• Encourage Stakeholders to promote AI principles and the guidelines outlined above,
via
▪ human oversight of AI systems amid; and
▪ discussion with different groups, to communicate harms and benefits,
incorporate feedback and raise awareness of the risk management process.
Petrolink supports ethical AI behaviour based on a risk-based approach, to conform with
present laws and the continual deployment of regulations and standards. All of these represent
components of continual improvement, including awareness of the AI dynamic ecosystem.

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 8
 Department(s): QM/HR
V1.0 - 22/Mar/2024
Security Classification: Internal

5. Further reading
Whilst the above sets out the basic principles of AI within the Petrolink Group, please also
refer to:
• our Human Rights Principles;
• our Privacy Notice (https://www.petrolink.com/privacy-notice/).
• our Code of Conduct, for more detail on areas such as Equal Opportunities, and other
topics touched on in this document.
• our Petrolink group Information Security Policies

6. References
Artificial Intelligence and Cognitive Science- 30th Irish Conference, AICS 2022 Munster, Ireland,
December 8–9, 2022 Revised Selected Papers.
Artificial Intelligence Risk Management Framework (AI RMF 1.0) - NIST AI 100-1, January 2023.
AI Standards & Assurance Roadmap - Action under ‘AI – Here for Good,’ the National Artificial
Intelligence Strategy for Ireland, NSAI, July 2023.
European Council, Press Release, "Artificial intelligence act: Council and Parliament strike a deal on
the first rules for AI in the world", 9 December 2023 ("European Council Press Release").
Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial
Intelligence, 30 October, 2023 (Press Briefing - The white house).
Four Principles of Explainable Artificial Intelligence, (NIST.IR.8312), September 2021.
Introducing the AI Safey Institute, UK Government, November 2023.
ISO 9001:2015 - Quality management systems -- Requirements.
ISO/IEC 22989:2022 Information technology — Artificial intelligence — Artificial intelligence concepts
and terminology.
ISO/IEC 23894:2023 - Information technology — Artificial intelligence — Guidance on risk management.
ISO 31000:2018 Risk management - Guidelines
ISO/IEC 42001:2023 Information technology — Artificial intelligence — Management system.
ISO 45001:2018 - Occupational health and safety management systems — Requirements with guidance
for use.
National artificial intelligence research and development strategic plan, NAIIO, May 2023.

Once printed/downloaded this document is considered an uncontrolled version

Security Classification: Internal

Location: Intranet
Path: HOME\00 - Policies\
Petrolink AI Principles - Page 9