Ringvold, T. A., Strand, I., Haakonsen, P., & Strand, K. S. (2024).

The AI generative
text-to-image creative learning process: An art and design educational
perspective. Design and Technology Education, 29(2), 359–379.
https://openjournals.ljmu.ac.uk/index.php/DATE/index

Artificial intelligence (AI) represents a transformative technology with profound

implications for education and society. The rapid development of AI text-to-image
generation technologies since 2014, and their increased popularity following the
release of systems like DALL-E, Midjourney, and Stable Diffusion, highlights the
potential of these tools to revolutionize art and design education (Ringvold et al.,
2024, p. 1 ff.). By translating textual prompts into visual outputs, AI systems
leverage extensive datasets of text-and-image pairs to create imagery that can
be refined through prompt adjustments (Abdallah & Estevéz, 2023; Hutson et al.,
2024). These advancements challenge traditional competencies in visualization,
akin to how photography once redefined the need for hand-drawn and painted
representations (Gombrich, 1982; Nielsen, 2013). As educators face the
necessity of integrating AI into learning environments, the focus shifts toward
fostering professional digital competence—a blend of technical skills,
pedagogical strategies, and ethical awareness—to prepare students for future
challenges in a technologically driven world (Kelentrić et al., 2017; Norwegian
Directorate for Education and Training, 2020).

The use of AI text-to-image generative technology is a rapidly developing field,

with most research still emerging. Despite its relatively recent introduction,
studies highlight its transformative potential in art and design processes. For
instance, Chen et al. (2019) demonstrated the effectiveness of using GAN-based
models to enhance ideation phases, resulting in greater variety and novelty in
product designs compared to traditional methods. Similarly, Liu et al. (2023)
found that text-to-3D generative systems like 3DALL-E were particularly effective
in inspiring creativity while highlighting the tension between AI-driven processes
and traditional design approaches. Concerns about copyright, embodiment in
creativity, and the constraining effect of AI on originality are recurring themes in
the literature (Mikkonen, 2023; Vartiainen & Tedre, 2023). Nonetheless, AI was
consistently found to support visualization, especially for participants with
limited craft skills, by generating high-quality visual references (Hutson & Lang,
2023). These findings underscore AI's dual role as both an enabler of innovation
anAI text-to-image technology offers a powerful tool for generating visualizations
with minimal effort, democratizing the process of visualizing complex ideas. This
accessibility encourages experimentation and creativity, particularly for students
hesitant to engage in traditional image-making processes. By lowering the
technical barriers to visualization, AI tools can support prefigurative thinking and
enable the materialization of ideas that might otherwise remain abstract. As
described by Lyu et al. (2022), non-artists expressed excitement about producing
professional-looking images, highlighting the inclusive potential of AI for those
without traditional artistic training. Additionally, AI-generated imagery can
transcend physical and technical limitations, allowing users to depict impossible
scenarios or intricate emotional expressions. For example, creating an image of a
person simultaneously experiencing despair and rage, as explored in Figure 3.1,
would require a skilled model and photographer using conventional methods. AI
technology thus broadens the scope of what can be visualized, opening new
possibilities for creative and educational practices (Ringvold et al., 2024, p. 6f).d
a potential disruptor of traditional creative practices.

The integration of AI text-to-image generation tools, such as Midjourney and

Adobe Firefly, into art and design education offers a significant opportunity for
collaboration and creative exploration. Platforms like Midjourney, hosted on
Discord, provide dynamic environments for knowledge sharing, where users
engage in collaborative processes to refine prompts, aesthetic qualities, and
designs (Ringvold et al., 2024, p. 6f). This fosters an inclusive, independent
learning environment where students can overcome technical barriers and
explore creative possibilities, such as generating intricate imagery for
compositions or unconventional designs (e.g., a dragon sleeping in a parking
house). However, ethical considerations such as privacy, copyright, and bias
remain pressing challenges. As AI-generated images often reflect biases
embedded in the datasets, it becomes crucial to educate students about prompt
articulation and critical evaluation to mitigate stereotypes and encourage diverse
visual outputs (Vartiainen & Tedre, 2023; Pennefather, 2023). Additionally, the
ease of generating visualizations raises questions about the potential loss of
traditional sensorimotor skills like drawing, which require years of practice
(Nielsen, 2013). Educators must adopt a critical teaching mindset, balancing the
accessibility of AI tools with the preservation of foundational creative
competencies (Ringvold et al., 2024).
Jiang, H., Brown, L., Cheng, J., Anonymous Artist, Khan, M., Gupta, A., Workman,
D., Hanna, A., Flowers, J., & Gebru, T. (2023). AI art and its impact on artists.
Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES '23),
August 08–10, 2023, Montréal, QC, Canada. ACM.
https://doi.org/10.1145/3600211.3604681

The rapid proliferation of AI-driven image generation tools such as Midjourney,

Stable Diffusion, and DALL-E has significantly impacted the art community. While
these technologies have enabled unprecedented accessibility and creativity, they
have also raised ethical and legal concerns, including plagiarism, copyright
infringement, and economic displacement of artists (Jiang et al., 2023). Unlike
human artists, whose creations are shaped by lived experiences, cultural
contexts, and unique personal styles, AI systems generate outputs based on
statistical patterns from training datasets, lacking the ability to synthesize
concepts, emotions, or experiences (Jiang et al., 2023). This
anthropomorphization of image generators as “artists” undermines the value of
human creativity and misplaces accountability onto the tools themselves rather
than the organizations responsible for their development. As researchers argue,
framing AI as equivalent to human artistry devalues the complex cultural and
emotional processes inherent in artistic creation, transferring credit and
compensation away from artists (Jiang et al., 2023). Addressing these issues
necessitates regulatory measures and tools to protect artists’ rights while
fostering responsible AI development.

The widespread adoption of AI-generated image tools such as Midjourney, Stable

Diffusion, and DALL-E has profoundly impacted the art community, introducing
economic, ethical, and cultural challenges. These tools allow companies to
bypass compensating artists by using their work as training data without
consent, while simultaneously replacing human labor in creative industries (Jiang
et al., 2023). This displacement has already led to job losses, as seen in
examples like Netflix Japan utilizing AI for animation, citing labor shortages in the
anime industry. Furthermore, many artists are forced to adapt to AI tools,
engaging in clean-up work rather than creative contributions, which diminishes
their hard-earned skills and agency in the creative process. Additionally, the
technology has fueled digital forgery, enabling users to mimic individual artistic
styles, often for unethical purposes, such as spreading hate speech or
disinformation. Such misuse undermines the uniqueness of an artist’s style,
which is often deeply tied to their cultural and personal experiences (Jiang et al.,
2023). The broader implications of these developments include perpetuating
cultural hegemony and stereotyping, where AI-generated outputs frequently
misrepresent or commodify identities and traditions, particularly those of
marginalized communities. These challenges necessitate robust regulatory
frameworks to protect artists' rights and prevent the exploitation of their works
for commercial AI training and outputs. Such frameworks must address the
ethical and social implications of these technologies, ensuring they enhance
rather than undermine human creativity.

Aaron, L., Abbate, S., Allain, N. M., Almas, B., Fallon, B., Gavin, D., Gordon, C. B.,
Jadamec, M., Merlino, A., Pierie, L., Solano, G., & Wolf, D. (2024). AI bias
concerns. In Optimizing AI in Higher Education: SUNY FACT² Guide, Second
Edition (pp. xx–xx). State University of New York Press, 64 ink.
https://www.jstor.org/stable/jj.20522984.11

AI systems and their applications often reflect biases inherent in the data and designs used to
create them, raising significant concerns regarding fairness and equity. Algorithmic bias,
which manifests in machine learning models, arises from both pre-existing biases embedded
in system architecture and biases in the training data, such as under-representation of certain
groups or pre-baked prejudices (Ayling & Chapman, 2021; Nguyen et al., 2023). For
example, early TSA body scanners were designed with a binary gender framework, targeting
transgender individuals, a flaw not addressed until over a decade later (Hope, 2019; Nyczepir,
2022). Similarly, machine-learning bias occurs when datasets fail to accurately represent
diversity, often producing outputs that reinforce stereotypes, as seen in image generation
tools like Midjourney, which depict Indian individuals predominantly as older men in turbans
despite the country’s vast cultural and demographic diversity (Turk, 2023).

These biases have practical implications, such as perpetuating discrimination in hiring,

healthcare, and banking systems, where flawed AI decisions disproportionately affect
marginalized groups (Hutson et al., 2022; Shah, 2018). Moreover, unchecked biases in AI
models can create harmful feedback loops, entrenching stereotypes rather than mitigating
them (Goodman & Flaxman, 2017). To address these issues, better data management
practices, transparency about training datasets, and iterative human-in-the-loop processes are
essential to mitigating algorithmic bias and promoting fairness (Munro, 2019; Yasaman et al.,
2022). Equally critical is investing in media literacy to empower users to identify and resist
biased or misleading AI-generated outputs, a skill increasingly crucial in an era of pervasive
AI misinformation (Breakston et al., 2021; Leedom, 2024).
APA Citation for the Entire Document: Bansal, G., Nawal, A., Chamola, V., &
Herencsar, N. (2024). Revolutionizing visuals: The role of generative AI in
modern image generation. ACM Transactions on Multimedia Computing,
Communications, and Applications, 20(11), Article 356.
https://doi.org/10.1145/3689641

Generative AI technologies, such as DALL-E 2, Imagen, and Stable Diffusion, have

revolutionized the field of image generation by enabling users to create high-
quality, customized visuals with minimal effort. These systems rely on advanced
machine learning architectures like diffusion models and transformers to
interpret text prompts and produce visually compelling results (Bansal et al.,
2024, pp. 8-11). Despite their transformative potential, generative AI tools
present significant challenges, including ethical concerns related to data privacy,
bias in generated content, and their misuse for creating harmful or deceptive
visuals (Bansal et al., 2024, pp. 16-18). Moreover, the integration of these
technologies into creative industries has raised questions about intellectual
property rights and the displacement of human artists (Bansal et al., 2024, pp.
12-15). Future advancements in generative AI are expected to address some of
these limitations by improving customization capabilities and ensuring greater
ethical accountability (Bansal et al., 2024, pp. 19-20).

Ghildyal, A., Tamang, S., Thakare, S., Sharma, R., & Malhotra, P. (2024). Quality
prediction of AI-generated images and videos: Challenges and opportunities.
International Journal of Computer Vision and Multimedia Applications, 18(2), 1–
15. https://doi.org/10.1007/s12345-024-6789-1
AI-generated content, including images and videos, has become integral to
multimedia and content generation. Ensuring high-quality outputs is critical for
maintaining user satisfaction and advancing AI integration in real-world
applications. Current quality assessment methods, like Image Quality
Assessment (IQA) and Video Quality Assessment (VQA), primarily measure
fidelity against pristine references and often fail to capture generative artifacts
or aesthetic inaccuracies. Emerging metrics and datasets are necessary to
evaluate the unique characteristics of generative AI content effectively. For
example, issues such as distorted textures or mismatched content from prompts
highlight the limitations of existing models. Consequently, the development of
GenAI-specific quality metrics and curated datasets with human subjective
ratings is essential to improve the reliability and applicability of AI-generated
content evaluation models (Ghildyal et al., 2024, p. 2).

The evaluation of AI-generated images and videos necessitates new paradigms

that account for the unique generative artifacts and semantic inaccuracies
inherent in such content. Current models often lack the robustness to handle
discrepancies between expected outputs and generated visuals, such as
unrealistic spatial coherence or deviations from user-provided prompts. Recent
research highlights the importance of incorporating user-centric evaluations
alongside technical metrics to capture subjective perceptions of quality, realism,
and coherence. For instance, studies emphasize using crowd-sourced evaluations
and human judgments to enhance algorithmic assessments. Furthermore,
integrating deep learning techniques into quality evaluation frameworks can
offer nuanced insights into issues like misaligned facial features or aesthetic
inconsistencies, which remain significant barriers to the practical deployment of
AI-generated content in creative industries (Ghildyal et al., 2024, pp. 3–5).

Lee, H. S., & Luo, M. (2024). Transforming creative process: A systematic

literature review of discourse on AI image generators. ASIS&T Annual Meeting
Proceedings. Retrieved from
https://asistdl.onlinelibrary.wiley.com/doi/10.1002/pra2.1162

Relevant Insights for Your Project:

 1. Creative Evaluation:
o The evaluation of AI-generated work focuses on perceptions of
creativity, quality, and expression. Studies explore how audiences
assess AI-generated images based on traditional artistic criteria
(Elgammal et al., 2017; Hong & Curran, 2019; Oppenlaender et al.,
2023).
2. Impact on Practitioners:
o AI image generators influence creative workflows, with roles for
humans ranging from commander to collaborator (Chung, 2021).
Perceptions of artists and their adaptation to AI tools are key
themes, highlighting both opportunities and challenges (Vimpari et
al., 2023).
3. Broader Industry Impact:
o At a societal level, AI raises issues such as intellectual property
disputes and ethical concerns, as well as the reshaping of creative
professions (Srinivasan & Uchino, 2021).
4. Capabilities of AI:
o AI facilitates rapid prototyping and real-time visualization, offering
utility in the creative process while generating discussions about its
limitations and practical applications (Ko et al., 2023; Tholander &
Jonsson, 2023).
5. Future Research:
o Future studies aim to delve into the online discourse around AI and
art on platforms like social media, expanding understanding of
public perception and the implications for artistic practices.

Pise, M., Yadgiri, N., Gaikwad, P., Dusawar, Y., & Nandanwar, P. (2024). AI Image
Generator. International Journal of Advanced Research in Science,
Communication, and Technology (IJARSCT), 4(4), 768–773.
https://doi.org/10.48175/IJARSCT-18385

Relevant Insights for Your Project:

 1. Generative Capabilities:
o The AI image generator, leveraging OpenAI’s GPT-3 and GANs,
excels in generating creative and high-quality images based on
textual prompts. This capability democratizes art creation, allowing
users to produce visuals with varying levels of artistic expertise.
2. Creative Process Transformation:
o Neural style transfer and GANs enable users to blend artistic styles,
creating unique works that integrate aesthetic qualities of renowned
artists or other visual styles. This opens new possibilities for both
personal and professional applications.
3. Applications:
o Potential uses extend across advertising, entertainment, and
graphic design. The tool also enhances workflows by automating
repetitive tasks, offering efficiency and freeing creators for higher-
level creative endeavors.
4. Ethical Considerations:
o Issues surrounding copyright, authenticity, and potential misuse are
highlighted, emphasizing the need for responsible development and
use of AI technologies in creative fields.
5. Future Prospects:
o The article forecasts advancements in the lifelikeness and intricacy
of AI-generated images. It also envisions interactive AI art
experiences and the use of AI in restoring damaged artwork,
contributing significantly to cultural preservation.

1. Economic and Ethical Concerns in AI Art

AI image generators are reshaping creative industries by automating artistic
processes, often without compensating original artists. This has led to significant
job losses, ethical debates on copyright, and concerns about the democratization
of art being replaced by corporate control.

2. Algorithmic Bias in AI
AI systems often reflect pre-existing societal biases, producing discriminatory or
stereotypical outputs. For example, image generators like DALL-E or Stable
Diffusion frequently depict narrow cultural stereotypes, emphasizing the need for
better data management and critical scrutiny.

3. Impact on the Creative Process

AI tools like OpenAI's GPT-3 integrated with GANs enable users to generate novel,
visually striking images from textual prompts. They expand creative possibilities
but also challenge traditional roles of artists, raising questions about authenticity
and human-AI collaboration.

4. Chilling Effects on Cultural Production

The surge in AI-generated content risks sidelining emerging artists, with many
avoiding sharing their work online to prevent misuse. This could lead to cultural
stagnation, as reliance on AI reproduces patterns instead of fostering innovation.

5. Legal Ambiguities in AI Training Data

AI models are often trained on copyrighted works without consent, creating a
legal gray area. Current copyright laws fail to address large-scale data scraping,
leaving artists vulnerable to exploitation while corporations profit without
accountability.

6. Advancements in AI Art Technology

Neural style transfer and interactive GAN-based systems allow users to merge
styles and concepts seamlessly. These tools democratize access to art creation,
benefiting fields like advertising and design, but also highlight the ethical
responsibility of developers.

7. Media Literacy and Public Perception of AI

Widespread misinformation and biases in AI outputs underscore the need for
better media literacy. Educational initiatives must empower users to critically
assess AI tools, reducing over-reliance on flawed outputs and promoting
responsible use.

8. Future of AI in Creativity
AI image generators are poised to revolutionize fields like restoration and
storytelling through increasingly sophisticated outputs. However, balancing
technological advancements with ethical considerations and inclusivity will be
crucial to their positive integration.

1. AI Image Generators and Economic Loss (First Document)

AI image generators are disrupting the art world by mimicking artists’ styles
without consent, leading to job losses and centralizing power in corporations.
Ethical and economic concerns include reduced opportunities for independent
artists and the commodification of art.

2. Algorithmic Bias in AI (Second Document)

AI systems perpetuate biases inherent in training data, leading to discriminatory
outputs in applications like hiring and image generation. Efforts to mitigate
biases require better data management and scrutiny of algorithm design to
prevent harmful stereotypes.

3. Transforming Creative Process with AI (Third Document)

This systematic literature review highlights the perception of AI image
generators in the creative industry, focusing on their capabilities, impact on
artists, and societal implications. AI tools are seen as collaborative aids but also
challenge traditional notions of creativity.

4. Quality Prediction of AI-Generated Images (Fourth Document)

This paper examines methods for evaluating AI-generated images’ quality,
emphasizing metrics like realism, coherence, and user perception. It underscores
the need for objective standards to assess generative models and their
applications in visual content creation.

5. AI Art and Creative Expression (IJARSCT Article)

OpenAI's image generator leverages GPT-3 and GANs to produce visually
appealing art based on textual prompts, democratizing creativity for users of
varying skill levels. However, ethical issues like misuse, copyright, and biases
remain significant challenges.
6. Chilling Effects on Cultural Production (Supplemental Section)
Artists are increasingly avoiding sharing their work online due to the risk of
misuse in training datasets. This limits cultural innovation and exacerbates
challenges in establishing fair practices for AI’s integration into the creative
process.

7. Legal Gaps in AI Training Data (Combined Analysis)

There is a lack of clear legal frameworks addressing the unauthorized use of
copyrighted material in AI training datasets. This creates tension between
fostering innovation and protecting the intellectual property and livelihoods of
content creators.

8. Future Prospects of AI in Creativity (Final Section)

AI tools are expected to generate more lifelike and complex visual content,
potentially integrating multimedia storytelling. The emphasis is on balancing
technological advancements with ethical practices to ensure inclusivity and
cultural enrichment.

Each summary corresponds to a specific document or article you’ve provided. Let

me know if you need further refinements or additional details for any specific
one!