The good news is that I’m a journalist, so I asked some artists, researchers, and art critics what they think of the aesthetics of AI art. First, I called Amelia Winger-Bearskin, an artist and professor at the University of Florida. Winger-Bearskin has cataloged several visual trends that she has noticed in recent AI art. She names a trend Nightmare Corp. – often illustrated by images conjured up by Google’s Deep Dream, an older generator released in 2015. He specializes in swirling, psychedelic imagery, like memories of a particularly harrowing LSD trip. “Prog rock influences, definitely,” she says. Another category Winger-Bearskin is exploring, which she calls Dada 3D, sounds a lot like the silly scenes I invoke when I play around with these generators. She describes it as “something like a surrealistic parlor game”.
Taxonomic trends aside, Winger-Bearskin has identified broader stylistic tics in these generators. She sees Western Disney-style animation and anime as obvious influences, as well as a tendency to treat whites as a default race — a result, she suspects, of these generators being trained on datasets heavily based on Western Disney-style animation, anime , and pictures of whites.
“The rhetoric of these companies is that you can do anything you can imagine. But of course, popular culture follows certain stereotypes and tropes.”
—Lev Manovich, professor at the City University of New York
Lev Manovich also pays close attention to this. The cultural theorist and professor at the City University of New York has been lurking on Midjourney’s Discord server for the past year, analyzing how people use the generator. After Midjourney released an update last fall, he saw some changes to what people were prompting the generator to do. For example, as people got better at depicting people realistically, requests for portraits of both men and women increased.
Digital artist Sam King first began following the AI art scene closely in 2021. Excited by what they saw, they began share shared her favorite images on social media and built a following as a curator as technology took off. They describe the earlier generation of generators as favoring “freaky, abstract stuff.” (These generators are known as Generative Adversarial Networks, or GAN. I’ve seen a few People Call this look rather uncreative GANism.)
King sees the latest generation of generators, so-called diffusion models, as stylistically independent. Just as oil paintings and watercolors produce noticeably different effects, GAN generators and diffusion generators produce noticeably different images. For example, if you want a more realistic rendition of Tony Soprano having a cappuccino with Shrek, the diffusion models are more likely to produce convincing results. “Theoretically, you can create all sorts of aesthetics with these machines,” they say. However, more realistic does not necessarily mean more stylistically varied. Like Winger-Bearskin, King sees Disney and anime influences appearing frequently, as well as comic book art.
“The rhetoric of these companies is that you can do anything you can imagine. It’s about that open border. But of course popular culture follows certain stereotypes and tropes,” says Manovich. He keeps seeing variations on multiple themes: “Fantasy, fairy tales, comics, video games.”
Author-created image using DALL-E and prompt “1970s prog rock album art”.
OpenAI via Kate Knibbs