In 2018, a portrait entitled Edmond de Belamy, from La Famille de Belamy sold for $432,500, nearly five times more than the buying price of the Andy Warhol or Roy Lichtenstein pieces that were present at the same auction. The sale sparked controversy, as this portrait was created entirely by the French AI art program Obvious, with an algorithm written in place of a human signature (Cohn, 2018).
Early forms of AI art have been around since 60 AD, when ancient Greek inventors were described as having designed machines capable of writing text, generating sounds, and playing music (Brett, 1954). The 1950s brought major advancements in the research and development of artificial intelligence, with the official founding of the AI academic discipline at Dartmouth in 1956 (Crevier, 1993) Since then, artists have experimented with AI for autonomous drawing and painting, 3D video, animated films, etc. with varying levels of success. While AI programs have been involved in art for centuries, the sale of Edmond de Belamy drew significant attention to the issue and marked a shift towards more mainstream concerns about the future of art and human creativity.
As technology advances, AI programs become increasingly human-like in ability. This innovation can open doors for increased productivity, new forms of creation, and wider accessibility for the art-making process. At the same time, it can also detract from the intrinsically human qualities that make art impactful, such as imagination and originality, and work against traditional human art methods. This paper will explore AI’s relationship with both human and artificial creativity, especially as it relates to art.
A significant factor in the discussion of creative potential for both AI and human artists is the factors that dictate creativity or a lack thereof. The objective nature of creativity makes it difficult to define, but Paul Jenkins (2023), filmmaker and founder of the art blog Brilliantio, names a few key components to creative thinking. “While there is no single definition of creativity in art,” he explains, “it usually involves the expression of emotions, ideas, and experiences through various artistic mediums.” Emotions are “intricately connected” to artistic creativity, and the translation of human emotion into meaningful pieces of art is “one of the key aspects of creative expression” (Jenkins, 2023). Imagination, the ability to generate ideas that are not necessarily based in reality, and originality, the uniqueness of those ideas compared to what has been previously created, are vital pieces of the creative process (Jenkins, 2023). These innately human aspects of creativity raise questions to whether AI can actually be creative.
A major piece of that discussion, whether AI can be creative, lies in the methodology of AI creation. Two of the leading AI art programs are Midjourney and OpenAI’s DALL-E, both of which provide text-to-image software. Midjourney “learns” to understand patterns, colors, shapes, and forms from a dataset of millions of images, combining a generator system and a discriminator system to mimic the human brain’s method of creation (Pryor, 2023). DALL-E is similarly trained, and its training procedure “allows DALL·E to not only generate an image from scratch, but also to regenerate any rectangular region of an existing image that extends to the bottom-right corner, in a way that is consistent with the text prompt” (“DALL·E: Creating images from text.” 2021). Both programs allow subsequent edits to be made to the initial picture generated from the prompt.
AI’s dependance on data in order to create calls into question AI’s capacity for human levels of creativity. As Bernard Marr (2023), world-renowned author and advisor specializing in business and AI, argued in a Forbes article, AI is not genuinely capable of original thought. “No matter how impressive a piece of computer-created poetry or artwork might be,” he explains, “it’s always built from blocks carved out of the data that’s used to train it” (Marr, 2023). While humans may develop ideas based on what they have previously seen, heard, read, etc., all of this information, comparable to the data analyzed by AI, is viewed through the lens of our own humanity—our unique feelings, perceptions, experiences, and personal beliefs (Marr, 2023). While algorithms may make connections between hundreds of data points in the creative process, they do it in a nonorganic, mechanized manner. Instead of the uniquely personal connections humans are capable of, AI can only determine what concepts are frequently connected together, producing only an emulation or illusion of human creativity (Marr, 2023). This does not render AI invaluable in the creative process, however, but instead contributes to AI’s use as a “digital extension of our ability to express ourselves, generate new ideas, and inspire an emotional response from our audience” (Marr, 2023).
While AI may not be capable of original thought itself, it can serve to enhance human originality. A study from the Gottlieb Duttweiler Institute, a German organization founded in 1962 to explore the future of technology and its impact on industry, suggests that AI can relieve us of monotonous and time-consuming tasks, allowing more time for activities that stimulate creativity (Bieser, 2022). At the same time, AI can hinder human creativity by pulling artists’ attention away from such activities, especially with the addictive algorithms of social media and other digital applications designed to consume as much of our time as possible. Such programs may keep artists from experiences of interactions that might support the development of creative ideas (Bieser, 2022).
Generative AI may also produce content that might inspire new ideas from human creators. Notable American artist David Salle experimented with AI programs, prompting the algorithm to create artwork based on a line from a poem that Salle was himself inspired by (Small, 2023). Throughout the trial, Salle critiqued the generated artworks just as he would with a human art student. Through this process, Salle was inspired by the combinations of his own style and the styles of famous artists that were reflected in what the algorithm produced (Small, 2023).
Other studies demonstrate that the influence of AI hinders the development of original ideas. A study by Dr. Eric Zhou of Harvard Medical School and Professor Dokyun Lee of Boston University concluded that while peak content novelty for artwork increases with the influence of AI, the average content novelty, defined by subject matter and unique concepts, and visual novelty, defined by uniqueness in stylistic choices, decreases overall (Zhou & Lee, 2024). These results suggest that AI can enhance originality in some but has an overall hindrance on novel content and artistry. Those who approach the artistic process with AI in similar ways to Salle, using AI as a tool to inspire creativity, benefit from the aided inspiration. On the other hand, those who use AI exclusively for new ideas without adding the collaboration of their own human perspective succeed only in producing less novel creations (Zhou & Lee, 2024).
Apart from the negative effects of AI on the productive development of ideas, AI is found to generally increase productivity in the creative process. AI enhanced human creative productivity by 25%, measured by the number of works that an artist creates in a month (Zhou & Lee, 2024). Additionally, the adoption of AI in the artistic process continues to increase productivity over time, with an initial 50% increase in productivity on average which doubles after the first month, gradually increasing until stabilizing at a level higher than initial productivity levels (Zhou & Lee, 2024). After his experiment, David Salle expressed how artistic algorithms can increase human artistic production. “As a painter you only have time to create a painting, but each painting contains within it all the painting you don’t have time to make,” he explains. “AI is a great tool because it allows me to see thousands of combinations — things that I would manually sift through in years are made with 5,000 versions in an hour” (Small, 2023). Collaboration with AI can speed up the process of art production, allowing for greater productivity and creative output.
Along with its impact on the human creative process, the rise of AI art has implications for the perception of human creativity, too. As AI advances, it can become increasingly difficult to differentiate between AI and human artwork, raising concerns that human artwork will be devalued by generative AI art. In 2023, German artist Boris Eldagsen won Sony’s World Photography Award with his AI generated photograph Pseudomnesia: The Electrician. After being announced as winner, BBC’s entertainment reporter Paul Glynn reports, Eldagsen revealed that his image was created with AI, explaining that “he used the picture to test the competition and to create a discussion about the future of photography” (Glynn, 2023). Similarly, in an indie book cover contest, controversy arose as the winning submission was revealed to have been created using Midjourney, despite the winning artist Sean Mauss insisting at first that he made the art himself (Sato, 2023). As with the case of Edmond de Belamy, the sale and appreciation of AI work at the same—or higher—level as human art exemplifies the complication of the art world that accompanies increased access to and artistic capacities of generative AI.
One of these concerns relates to a loss of integrity in art evaluation when AI is paired, unknowingly or not, with human artwork. After it was revealed that Mauss’s book cover was created with AI, author Mark Lawrence, the organizer of the annual competition, announced that there would not be a competition going forward, explaining that he’s “uninterested in litigating future debates about whether art is human or machine-made” (Sato, 2023). Because AI art is so often indistinguishable from human-made artworks, artists and art critics are concerned about the possibility of AI art replacing human art commercially or human art being devalued as it becomes increasingly difficult to confirm that human-made art was not actually created by AI (Sato, 2023).
Despite fears that human artists are disadvantaged by AI art programs, a preexisting bias against AI can protect human artists from the risk that AI could replace them. A study conducted by Drs. C. Blaine Horton Jr., Sheena S. Iyengar, and Michael W. White (2023) of Columbia Business School explores how bias against AI influences how artworks are valued, concluding that AI-labeled art was consistently valued lower than human-labeled art, that human-labeled art was consistently viewed as more creative, and that this creativity rating increased when compared with AI-labeled art. This bias offers opportunity for human artists, who may increase perceptions of creativity and value by directly comparing their art to AI-made artworks.
Another solution to protect the value of human creative work is to judge AI as a separate entity in itself rather than alongside completely human-made work. After his photograph won, Boris Eldagsen thanked the judges for “selecting [his] image and making this a historic moment,” but refused to accept the award. “AI images and photography should not compete with each other in an award like this,” he explained. “They are different entities. AI is not photography. Therefore I will not accept the award” (Eldagsen, 2023). When Lawrence announced that he would no longer host his book cover contest, he offered his suggestion for the future of AI in similar competitions, tweeting, “I think it needs to be a separate contest, organized by someone with the necessary expertise and the appetite for controversy. That’s not me” (Sato, 2023).
Others take this idea one step further, suggesting that AI art should not only be judged separately but should be identified as a different category of art altogether. Dr. Marian Mazzone, professor at the Department of Art & Architectural History at the College of Charleston, and Ahmed Elgammal, professor at the Department of Computer Science at Rutgers University, propose that AI algorithms should be viewed as more than just artistic tools and closer to an entirely new medium (Mazzone & Elgammal, 2019). “The word medium in the art world indicates far more than a tool,” they explain. “A medium includes not only the tools used (brush, oil paint, turpentine, canvas, etc.) but also the range of possibilities and limitations inherent to the conditions of creation in that area of art” (Mazzone & Elgammal, 2019). With this view, AI does not work against human creativity but rather acts as a new vessel for creative work.
As generative AI technology continues to progress, algorithms will have an ever-increasing role in the art world and, in a broader context, all areas of human creativity. As artists and audiences alike grapple with these changes, the challenge lies in embracing AI’s capabilities while honoring the irreplaceable value of human creativity. While debate remains as to whether AI is a greater help or hindrance to human creativity and its perception, AI’s impact on the world of creativity, artistic and otherwise, is significant and undeniable.
Works Cited
Bieser, J. (2022). Creative through AI.https://gdi.ch/en/publications/studies/creative-through-ai
Brett, G. (1954). The automata in the byzantine “throne of solomon.”speculum, 29,3 no https://www.journals.uchicago.edu/doi/10.2307/2846790
Cohn, G. (2018 October 25). AI art at christie’s sells for $432,500. The New York Times https://www.nytimes.com/2018/10/25/arts/design/ai-art-sold-christies.html
DALL·E: Creating images from text. (2021). https://openai.com/index/dall-e/
Eldagsen, B. (2023,March 14). Sony world photography awards..boris eldagsen. https://www.eldagsen.com/sony-world-photography-awards-2023/
Glynn, P. (2023, April 17). Sony world photography award 2023: Winner refuses award after revealing AI creation. BBC. https://www.bbc.com/news/entertainment-arts-65296763
Horton Jr, C. B., White, M. W., & Iyengar, S. S. (2023). Bias against AI art can enhance perceptions of human creativity. Scientific Reports, 13(1), 1–14. https://doi.org/10.1038/s41598-023-45202-3
Jenkins, P. (2023, July 25). What is creativity in art: Unlocking the secrets of artistic expression. brilliantio. https://brilliantio.com/what-is-creativity-in-art/
Marr, B. (2023, March 27). The intersection of AI and human creativity: Can machines really be creative? Forbes.
Mazzone, M., & Elgammal, A. (2019). Art, creativity, and the potential of artificial intelligence. Arts, 8(1), 26. https://doi.org/10.3390/arts8010026
Pryor, J. J. (2023,). How does midjourney work and function? A beginner’s overview. https://threwthelookingglass.com/how-does-midjourney-work-and-function/
Sato, M. (2023). How AI art killed an indie book cover contest. The Verge
https://www.theverge.com/2023/6/9/23752354/ai-spfbo-cover-art-contest-midjourney-c larkesworld
Small, Z. (2023, September, 22). A.I. excels at making bad art. can an artist teach it to create something good? The New York Times
https://www.nytimes.com/interactive/2023/09/22/arts/design/david-salle-ai.html
Turning, A. (1993). Computing machinery and intelligence
Zhou, E., & Lee, D. (2024). Generative artificial intelligence, human creativity, and art.
PNAS Nexus, 3(3). https://doi.org/10.1093/pnasnexus/pgae052
