As generative AI becomes increasingly ubiquitous and advanced, institutions at various levels are reevaluating their policies on how to work with this tool, or questioning if they should at all. Since President Carmen Twillie Ambar’s announcement of the Year of AI Exploration at Oberlin this fall, the Studio Art faculty, staff, and students have raised questions and concerns about what the emergence of AI means for their field. Will AI replace artists? What are the ethical and environmental implications? Should we proceed with caution or dive in with full force?
The first AI software in the world of fine arts dates back to the 1970s, when abstract painter Harold Cohen sought new ways of creating art using computers. Cohen debuted a program called AARON that, unlike today’s generative AI, would draw images from scratch. However, it was only recently that AI became mainstream due to its increased accessibility and convenience. With OpenAI’s 2021 announcement of DALL-E, anyone with access to the internet is now able to generate images from a text prompt, not just tech-savvy researchers.
Assistant Professor of Sculpture Zé Kielwagen noted that the pattern of fear and uncertainty surrounding new technology’s impact on artists has been around long before the recent popularization of AI.
“When photography first became popular, painters were paying a lot of attention to it,” Kielwagen said. “Photography brought the cost of image making down and therefore threatened well-established careers at the time. But at the end of the day … [painters] didn’t have the burden of having to represent the world anymore, so they were free to [explore] … abstraction and surrealism.”
While the advancement of photography eventually had a largely positive impact on painting as a discipline, it is not guaranteed that AI will have the same effect. In fact, many artists are infuriated about the legal authorship implications of generative AI. Unlike Cohen’s AARON, today’s generative AI does not build from scratch; it pulls from a database of work already established by artists. While tech companies profit off of this, the original creators of the work are not receiving the same benefit. Kielwagen commented on this issue, offering an alternative perspective.
“There’s a copyright concern [for artists whose] images are being used and processed [through AI] to make someone else’s work,” he said. “The creator of the original work is not compensated for that. But then what about collages? … You’re also appropriating the work of that photographer, that designer, the person who prepared that magazine. I agree that we have to think about [generative AI], but we have to think about everything else with the same rigor.”
As Kielwagen expressed, the discussion surrounding AI — especially within artistic spaces — is incredibly nuanced. Eileen Maxson, visiting assistant professor of Integrated Media, echoed the legal concerns about AI, but she emphasized her belief that the primary issue is something else entirely. As interest in AI increases, researchers aim to develop the fastest, most efficient routes for software to take, eliminating the time it would take for slower, traditional forms of art-making. Maxson suggests this may be detrimental to the real value of art.
“There are ethical and legal questions about the models that were used to create AI that are going to potentially take away opportunities for illustrators to have a space to publish their work as well as support themselves,” she said. “I think the really problematic thing is the idea of human expression being crowded out by AI expression. … Part of what’s enjoyable about the creative process is all the detours and connections that you make.”
Michael Boyd Roman, assistant professor of Studio Art and Africana Studies, expands on the value of the creative somersaults that art inspires. To Roman, it is worthwhile to skip the shortcut of generative AI and embrace the idiosyncratic elements of the artistic process.
“One of the things I love about teaching art, comics, and drawing is that we can look at the exact same object and have all of these different views of it,” he said. “There’s something unique about it coming through the sensory input of your eyes, being translated in your brain, and then going through all the mechanics of your shoulder, elbow, wrist to fingers, through the mark-making tool and onto a surface. Each of those points is a different process for every person. I don’t want someone else’s process interfering with mine; AI is trained on other people’s art.”
Another area of concern within the art community is what AI means for their job security. Jamie Jacobs Overstreet, program coordinator for the Studio Art and Art History departments, expressed this unease.
“As a union member and supporter, I have some concerns about job losses and displacement from [AI],” she said. “But again, we all have to engage with it more to be able to fully flesh out our opinions and try to head off concerns with solutions that make us uniquely human.”
On the other hand, Myles Dunigan, 2D studio manager and lecturer, asserted that AI may not be as large of a threat as some artists may believe or be fearful of.
“There’s probably more of a risk in design-based disciplines with a fast turnaround,” he said. “But from a fine art perspective, I don’t know that AI is going to have the same level of craftsmanship or complexity that we would train people to think about and make in a fine arts setting. … [For example,] AI music has a lot of tropes that it repeats in a bland and predictable way.”
The generative AI that we know today is still in its early years. As it rapidly develops, Ling-Lin Ku, assistant professor of Studio Art, suggests that our ways of thinking about art and technology will follow.
“One of the things for us to think [about is] what art can do,” she said. “What’s the essence, what’s the nature of art? AI, just like a lot of technology in the past, forces us to think about what art is. [It promotes a] very high form of reflection of what human beings think and what creative decisions they make. … [Art will] continue to change, take different forms, answer different questions, and ask different questions.”
Another sizable concern surrounding AI is its environmental implications. With hefty water and electricity requirements, AI’s carbon footprint is of no small scale. In fact, the International Energy Agency reported that one request made through ChatGPT consumes nearly 10 times the electricity of a Google search.
“[The environmental impact] is the chief concern,” Roman said. “The idea that these systems are having real world consequences for the people that are unfortunate enough to live close to the real world locations of these things, which are often highly racialized, is an issue … My luxury should not come at the expense of someone else’s health and well-being … If we can’t produce it cleanly … or ethically, then we shouldn’t be producing it. But there’s no money in that.”
With ongoing conversations about the usage of AI art, the Studio Art department encourages the artist community to challenge perspectives and voice their opinions.
“I know I sound very definitive in my statements here, but I’m always open to a conversation,” Roman said. “If you have questions, talk to your faculty. If you differ in opinion, I want to hear about that. I’m not above having my mind changed.”
