Cindy Sherman, the artist known for her uncanny portrait photographs, once described her process as summoning. Clad in wigs and costumes, and daubed with exaggerated makeup, Sherman has spent decades photographing herself as other people.
“I want there to be hints of narrative everywhere in the image so that people can make up their own stories about them,” she said, in 2016. That year, the Guardian described Sherman as “the original shape-shifting selfie queen”, before enumerating a few of her many guises:
20 kinds of matinee starlet
Hitchcock lead
pneumatic Monroe
crime-scene corpse
old master muse
Republican wife
I don’t remember when I first encountered Sherman’s work, but it was around the time I entered art school, if not before. As an aspiring portrait photographer and classic film fan, Sherman’s Film Stills caught my attention, and undoubtedly influenced some of the assignments I submitted during the first couple of years in my BFA program.
Even then, I was sensitive to Sherman’s themes—of visibility, youth, and aging. I grew up in the early-mid 2000s, when femininity was typified by Paris Hilton, Britney Spears, Lindsay Lohan, and Kim Kardashian. Famous party girls, who, if they weren’t buxom and blonde, were thin or rich. (In Hilton’s case, thin and rich.)
Over her long career, Cindy Sherman has role-played similar archetypes, and has, as the Guardian noted, anticipated the selfie culture now embodied by the Kardashians. At the start of her career, Sherman might only have dreamed of the augmented reality tools now available to us on Instagram: filters that can change our eye colour, apply makeup, hats or sunglasses, and move with us in a cascade of glitter, lashes, and … whatever this is … as we pout and vamp for the camera.
In recent years, Sherman has shared her work on Instagram, using it as its own medium. (In 2017, the Guardian wrote, quaintly, “…the work seems to be made specifically for Instagram, and not as a physical work for a future show in a gallery.”)
Following an injury, Sherman shared portraits she’d created with FaceTune and other apps. “I was kind of just lying around with nothing to do and playing with my phone,” she told W magazine. Even then, Sherman was looking ahead. “I feel like I’m looking for something else—another place to take it to,” she added.
So, let’s take it somewhere else, shall we?
Before we do that, I want to go back to Sherman’s own description of her work as a summoning.
It sounds esoteric, like she’s engaged in a process of divination—scrying and scrolling the ley lines of her own form, dowsing the pale blue veins on the backs of her hands, miming and mining the body at once.
As I see it, Sherman’s work is not only a summoning, but an extraction.
“[I was] working my way out of the photographs: I went from being a reflection in a mirror to a fuzzy figure in the background to a body lying beside ants and fake blood,” Sherman told the Guardian.
In her “disasters and fairytales” (1985-89), she exhumed all kinds of horror using fake limbs and plastic boobs; subsequently, things got even more grotesque, with lavish photographs featuring fake vomit and rotting food. (That period was, she says now, a direct challenge to the art market that had appropriated her as its latest squeeze. “I was like, do they really like the work? Well, let’s see them put this over the dining table – I had fun making those pictures.”)
What do you do when something festers? You take it out.
What do you do when something beckons? You let it in.
To summon, after all, is to find. To remember. To call. To invite, invoke, or coax.
Fittingly, text-to-image generators mimic this process; conjuring one-of-a-kind images from written prompts. I wrote about Stable Diffusion’s recent upgrade in my previous post, and since then, they’ve added a new way to work with the app: you can now enter a negative prompt. In other words, you can tell Stable Diffusion what not to generate. In libraries (I’m a librarian), we call this Boolean searching, where the qualifiers AND, OR or NOT are added to a search string for greater precision.
To be honest, I haven’t tried this feature yet. I’ve been experimenting with something else. Stable Diffusion Prism is an image-to-text generator that works by tapping into OpenAI’s CLIP; a neural network trained on text and images found online. (I know, I know: none of these words are in the Bible; bear with me.)
I wondered how Artificial Intelligence might interpret Cindy Sherman’s style, and how it would render her body of work. As it turns out, the AI knows exactly how to mimic the ultimate mimic artist. The details—hands and eyes—are a bit wonky, but then again, so are Shermans’s, especially in her Instagram era.
Nonetheless, the results are actually kind of brilliant. Stable Diffusion has synthesized so many of Sherman’s tropes while generating an infinite variety of new forms modelled on her aesthetics.
Is Cindy Sherman working with Artificial Intelligence? If not, she should be.
EDIT: DEC 17/22 Sherman appears to be working with AI! The post embedded here is captioned “AI”.
The following images were rendered by Stable Diffusion in the style of Cindy Sherman, using simple prompts like “Cindy Sherman” and “Cindy Sherman Film Stills”.
The captions below each image were generated by Stable Diffusion Prism, based on the images generated by Stable Diffusion. Any biases or errors are the result of the AI’s own biases and limitations.
“a close up of a person wearing a red shirt, a character portrait, by Cindy Sherman, trending on deviantart, neo-primitivism, grimes - book 1 album cover, she is about 7 0 years old, dramatic angle, trending on tiktok, portrait of the old necromancer, long yellow hair”
“a woman with blonde hair wearing a white dress, a character portrait, by Cindy Sherman, hyperrealism, from movie bladerunner, renaissance style painting, mid-transformation, 1 9 8 6 movie screenshot, beautiful jewish woman, pale and sickly, unmasked, 2 0 5 0 s, face - up”
“a woman in a white dress with a big hairdo, a flemish Baroque, by Cindy Sherman, behance, portrait of lady gaga, artforum, realistic restored face, berghain, toiletpaper magazine”
“a black and white photo of a woman smoking a cigarette, a character portrait, by Cindy Sherman, behance, art photography, looking in mirror at older self, the girl is scared, blonde woman, mirrored, pinhole camera, sexy movie photo, uma thurman, canon eos ef 5 0 mm”
“a black and white photo of a woman with her eyes closed, a photorealistic painting, by Cindy Sherman, hyperrealism, wearing modern headphone, moody : : wes anderson, facing left, lois greenfield, featured on vimeo, with a halo of unkempt hair, dan witz, zdzidaw”
“a close up of a person with a ring on their finger, poster art, by Cindy Sherman, massurrealism, smoke coming out of her mouth, georgia o'keefe, purple volumetric lighting, jan saudek, ███beautiful face███”
“a black and white photo of a woman with a hood on, a character portrait, by Cindy Sherman, bauhaus, medium close - up ( mcu ), frequency indie album cover, anthropomorphic mare, photorealistic artstyle, age 2 0, movie still of a tired, early 20s”
“a woman with a surprised look on her face, a character portrait, by Cindy Sherman, featured on cg society, precisionism, photographic style of avedon, muscular ultraviolent woman, coloured polaroid photograph, frank sedlacek, older woman, crossbreed, bioart, on a young beautiful woman neck”
What stands out to you here? For me, it’s how readily Stable Diffusion Prism performs the kind of gaze Sherman critiques in her work. Look at its keywords: beautiful, scared, sexy. There’s an emphasis on her age, too, as well as vulnerable features like her neck and mouth. A few other descriptors—muscular, ultraviolent, tired, blonde—evoke Sherman’s sense of unease.
When Sherman said she hoped people would make up their own stories about her work, would she have imagined that it wouldn’t be long before Artificial Intelligence made up its own stories about her work? Share your thoughts in the comments, and subscribe below for more from DALL·Eance; art criticism for the Extremely Online.