Cindy Sherman is ‘Experimenting’ With AI and Not Everyone’s Happy

Cindy Sherman AI portraits
Cindy Sherman’s recent AI self-portraits, but she insists they are not for exhibition.

Photographer Cindy Sherman, best known for her off-the-wall self-portraits, has raised eyebrows recently by experimenting with artificial intelligence (AI).

Sherman began posting AI-generated selfies this year and tells The Art Newsaper that she has been using the Lensa app.

“I am really curious about AI and what it does to images. I was playing around with one of these face-tuning apps, the ones that are designed to make you look better,” she writes in the Finanical Times.

“The way it works is that you load in selfies, and it creates multiple versions of you. Only I was making myself look weird, or just different. So I put those selfies into one of the AI programs — it creates various “types” of you; anime, cosmic, apocalyptic, etc. So it was a sort of double manipulation.”

Not Everyone is Down With AI

However, when Sherman shared her creations on Instagram she faced a furious backlash. Going so far as to leave a comment under one of her posts, “People, I’m just fooling around, sheesh, it’s not my ‘new’ work. Do you really think I’d post new work before exhibiting it?”

Explaining herself further in the FT she says that is just “experimenting” and likens Instagram to a “sketchbook.” But she does admit to being curious about AI, calling it a distortion and a mirror.

“The new images I’ve made for the show in Zürich were partly inspired by some of the ways I’ve been playing around with AI,” she says.

“The selfie app produced all these weird things, big heads growing out of the ground, a hand with six fingers, two hands coming out of one arm. Really strange stuff. That was part of what was in the back of my mind. Even though they’re created using old-fashioned collage, I wanted it to seem like some really strange distortion was going on with these faces.”

Sherman explains that Lensa works by feeding it a batch of selfies which the AI makes avatars out of.

“Generally, I think they are trying to make really attractive avatars of your face. But because the images that I’ve given them are these altered images, the results are just so much more surprising,” she tells The Art Newspaper.

“Some of the characters seem to have two hands growing out of one arm or the face seems kind of chopped up. That partially inspired me for this new body of work. People were responding to my Instagram posts saying, ‘Your other work was so much better,’ but I don’t consider [Instagram posts to be art] works. It’s just fun, I’m playing around.”

The crux of peoples’ complaints about Sherman’s use of AI is how the technology was built. Billions of images were scraped from the internet to build the alogorthms that produce often-stunning synthetic imagery. But of course, the training data includes billions of copyrighted photos whose owners were not asked if they consent to such usage.