Skip to Main Content

WM: Blurred Lines: How AI Is Changing Art

Artificial Intelligence (AI) has rapidly made its mark on nearly every industry in the past decade, and the arts are no exception. As the technology becomes more widely available to the public, questions grow about the impact it will have on the future of creative work.

 

WHEN HE’S NOT IN THE CLASSROOM helping creative minds work on their next big project, Damon Mohl stays busy crafting his own works. From filmmaking and set design to painting and drawing, the Wabash art professor likes to keep his hands and mind busy.

During semester break while working on a short film, Mohl thought he’d try experimenting with something new: artificial intelligence (AI). 

He opened Adobe Firefly, a suite of generative AI tools trained to recognize connections between text and images to allow users to generate imagery using text prompts. 

“I was interested in the distortion and strangeness that a computer algorithm could come up with images using a simple prompt written by the user,” Mohl explains. “I could see all of that working as a tool.

“If I am working on an animation and I want to duplicate an object 300 times, I could take the time and do that by hand or I could use AI to fill in that space,” he says. “I wouldn’t have any problem with that—it’s the same thing as using Adobe’s clone stamp or all these other utilities that allow you to create images.”

Like many other artists exploring the technology, Mohl did have some initial hesitations using AI in his work.

“One of the dangers for a visual artist would be if you’re taking out the actual creative process of an idea. If there’s no longer this sort of human struggle to create or explore something, then there’s no real point in doing it,” says Mohl. “It’s a complicated topic for sure because every artist will have a different line as far as how they would interact or incorporate the technology into their practice.

“Some might not have any need or use for it. Some might be able to use it in an interesting way specifically as a supplemental tool. And others will use it in probably the worst way where they’re letting the system make all the choices and do all the work for them.”

 

IN ORDER TO UNDERSTAND the impacts of AI on art specifically, it’s important to first understand what exactly the technology is.

“AI has a funny history because there have been blips of activity and then things die off and go out of vogue. What we’re seeing is more than a blip,” says Colin McKinney, a Wabash mathematics and computer science professor, whose first introduction to artificial intelligence came when he was in middle school. He found a book on the subject and was eager to learn more.

Adobe Firefly image of cat playing guitar in style of Van Gogh“In some sense, it’s not new, but certain technological advancements that have happened in the last 10 to 20 years have really accelerated the scope of AI and the extent to which we can actually implement it.”

AI image generators, like DALL-E 2, Midjourney, and Stable Diffusion, are “trained” on an extensive amount of data, such as large databases of images, videos, or writing. The generators learn, for example, what a cat is, or a guitar, or Van Gogh’s style of art. Then, when a user asks for an image of a guitar-playing cat by Van Gogh, a new image is created based on the relevant information in the generator’s dataset.

 

IN SOME CASES, it can be relatively easy to detect when something was created using AI. 

McKinney referenced ChatGPT, a free-to-use AI system that is trained to follow an instruction in a prompt and provide a detailed response. The application is often used as a tool or sounding board to help users write speeches or essays.

“I played with it maybe a year and a half ago when ChatGPT entered the common vocabulary and within five minutes I could figure out its extreme limitations in terms of making things up that don’t exist, called hallucinations, and wrong facts,” says McKinney.

“There have been high-profile cases of people misusing it recently. Michael Cohen, who was an attorney for President Donald Trump, and his team submitted some court documents that referenced nonexistent court cases, which is pretty clear evidence that they were using some sort of ChatGPT or other generative system to produce their work,” he says. “That tends to be what busts people. The text itself might be fine, but the references are bogus.

“There are other instances, specifically within the art world, where it can be more challenging to catch whether content was human-computer generated or machine-computer generated,” says McKinney.

Research from Bowling Green State University published in the journal Psychology of Aesthetics, Creativity, and the Arts in 2023 found that when shown a mix of AI and human visual art, participants correctly identified the source a little more than half of the time and their guesses also exhibited a low level of confidence.

 

AT FIRST, AI ART-GENERATING programs struggled with creating images, especially of people. They had eyes looking in different directions or had hands with extra fingers. 

But as these applications advanced, the tools have become better, making it easier to create hyperrealistic faces, which has led to the rapid increase in deepfakes. The term “deepfake” refers to an image, a video, or an audio recording that has been digitally altered to mimic a realistic human experience. This includes superimposing human features on another person’s body, manipulating sounds, and creating videos that did not occur but seem authentic.

The term came about in 2017 when a Reddit moderator created a subreddit called “deepfakes” and began posting videos that used AI to insert celebrities’ likenesses into existing pornographic videos. Pop superstar Taylor Swift is one of the latest victims of deepfakes after AI-generated nude images of her were produced and distributed on social media without her consent. 

Other examples of deepfakes include a video showing soccer star David Beckham speaking fluently in nine languages and an image of Pope Francis wearing a stylish white puffer jacket. Neither of these events happened.

“There are two sides to this,” says McKinney. “(As an example) think of images of the American Confederacy. Those images could be used essentially for educational purposes in a historical movie or artist’s reconception of the surrender at Appomattox. Or those images could be used for really terrible hate speech purposes.

“The courts have long held that there are reasonable exceptions to First Amendment privilege, and maybe this needs to be a new exception,” he continues. “But these are really, really tricky issues that are going to take decades to work through courts, legislation, or both because any law that gets passed is immediately going to get challenged by somebody.”

 

AS THE TECHNOLOGY continues to evolve, Chris Rozzi ’90 believes it’s important for artists to educate themselves on AI and the potential impacts it could have on their personal and professional livelihoods. 

Rozzi is the lead experience designer at Körber Supply Chain, a software consulting company, with 30 years in visual design and 14 years in user experience design. The Wabash art major says he depends on generative AI daily in his role with Körber. The technology helps him in his research by taking large amounts of data and turning it into proactive optimizations for customers.

In his free time, he also uses AI art generators like Firefly as an inspirational tool to help spark new ideas for drawings.

“AI will continue to be such a transformational tool that you can’t just not pay attention to it,” says Rozzi. “No matter what field you pursue—whether you want to be a traditional painter or an illustrator—you cannot escape technology. You need to embrace it. It will only be to your detriment if you don’t at least become aware of it, step into it, and learn how it can help in regard to your art or the work you are doing.”

Mohl says he wants his students, especially those who hope to continue pursuing the arts after Wabash, to at least be aware of AI’s possibilities.

“This concept is not anything new. The invention of the camera changed painting,” says Mohl. “There’s always going to be new technology that artists will work with and adapt to, find interesting ways to use, or reject all together.

“As an artist, you discover what you like and what you don’t like through making,” he says. “Whether or not AI is a part of that process for you, everyone has to decide that for themselves. My hope is that creativity never dies. The whole point of art is the struggle in exploration of something. If you reach a point in your art where you’re no longer in control of your ideas or what you’re doing, there’s no real reason to do it.”

 

ANOTHER CONCERN AMONG ARTISTS is that AI generators infringe on copyrighted materials and could ultimately undercut the market for their work.

AI art generators rely on human-made images for training data, Rozzi says, and these “scrape” billions of images from the internet, which can include countless works by digital artists who never gave their consent. 

Rozzi encourages artists to practice the lessons that a liberal arts education teaches when working with AI. 

“Be a critical thinker, never stop asking better questions, and keep learning,” says Rozzi. “There’s so much happening with AI—ethically, legally, creatively—that we don’t have clear answers to and probably won’t anytime soon. That doesn’t mean we should be afraid of it. How we choose to respond and interact with the technology is ultimately up to us.”

Back to Top