By Malia Mendez
Art Versus Artificial: AI-Assisted Artists Vie for Critical Recognition in San Francisco Exhibition
August Kamp never thought her art would make it into a gallery.
Kamp is a generative artist: someone who creates artworks with the help of computer software. Therefore, people have often told her that she is not a "real" artist. As a trans woman, Kamp said, this insult is familiar.
But when Steven Sacks, director of bitforms gallery in San Francisco, decided to curate an AI art show, Kamp was just the kind of artist he was looking for.
“I wanted to curate a show that basically had a wide interpretation of how AI is being used… to have a dialogue with the person who made their first great DALL-E piece a month ago versus someone who’s been practicing [digital art for] 20 or 30 years,” Sacks said.
The resultant exhibition, “Artificial Imagination,” opened at bitforms on Oct. 26 — one month after Kamp released her most-viewed AI artwork, an augmented-reality reimagination of Johannes Vermeer’s “Girl with a Pearl Earring” — and closed just before the new year.
When Kamp was invited to show her work as part of “Artificial Imagination,” she attributed her feeling unfit for the gig to first-time nerves. But after meeting the seven other participating artists at the exhibition’s opening, she realized several shared her puzzlement.
“With a couple of exceptions, we were all gathered together talking and being like, ‘I never thought my art would be in a gallery,’” Kamp said, especially given that, in the media, “the AI community is represented as these people who don’t care, are happy to steal.”
Tech companies like OpenAI are starting to release artificial intelligence tools like DALL·E, an algorithmic image generator, to the public at a greater frequency. As a result, those who incorporate such tools into their art practice — and are now being shown at experimental galleries like bitforms — are increasingly being accused of at best unoriginality, and at worst copyright infringement.
In early November, programmer and lawyer Matthew Butterick filed a class action lawsuit against Microsoft, its subsidiary GitHub, and its business partner OpenAI, alleging that the companies’ AI-powered coding assistant Copilot was trained on stolen code. Because DALL·E, like Copilot, draws from web-scraped images that may be copyright-protected, Sacks said he would not be surprised to see similar legal action in the art world.
Piero Scaruffi, a famed cultural critic and self-declared “provocateur” who has hosted lectures at the intersection of art and science at the University of San Francisco and Stanford University since 2008, said that he was glad when he read about Butterick’s lawsuit.
“Some of the major news outlets made fun of it as if it was just a funny thing,” he said. “In my opinion, it’s very serious.”
While the music industry is consistently flooded with fair use cases, Scaruffi said, the public has yet to mobilize against DALL-E’s web-scraping model.
“All you have to do is copy 10 seconds of somebody’s melody, and you get in trouble,” Scaruffi said. “So DALL-E learned based on how many images? One billion. And it doesn’t owe anything to all the people who created those images? This sounds a little unfair.”
In light of Butterick’s lawsuit, Kamp said many of her generative artist peers have recently been turned down by galleries outright because they see AI art as a “copyright tangle.” Others she knows are suffering under a barrage of online abuse, some of which is coming from other artists.
“If you say, they stole your work, then so have you,” one generative artist wrote in a tweet addressing the more “traditional” artists discrediting her work. “You went to the gallery, you scrolled down through Artstation, you photobashed. If this comparison confuses you — educate yourself about how [AI] works.”
Kamp said almost every artist she knows started out by imitating someone whose work they loved. Imitation, she said, is inherent to artists’ growth and development; no one is truly original.
“I think that if an artist owned a style, Star Wars would never happen, because Star Wars ripped off a bunch of other stuff and brilliantly ripped them all off at once,” Kamp said. “That’s how creative people do stuff.”
Embracing AI art is not about disenfranchising "traditional" artists, Kamp said, but providing everyone the tools to express their creativity — especially those who might not otherwise have the time, money, or physical ability to do so.
For Ellie Pritts, another artist showing work at “Artificial Imagination,” AI was the reason she was able to keep making art after a debilitating neurological disorder left her unable to draw or play cello for an extended time, both of which she’d been doing since she was a child.
“There is a lot of ableism in the way people talk about how AI art is easy,” Pritts said. “You take a person like me, and I cannot physically do this anymore the way it used to be done. Does that mean what I’m making is less valuable? I would say no.”
Pritts knew something was wrong when she began experiencing mobility issues and temporary paralysis as a teenager, but she wasn’t diagnosed until she could finally afford health care this year. In August, she broke down after trying and failing to draw a portrait of her partner.
“It was a very sad thing for me, but then I discovered AI, and was able to basically take the art that I have made in the course of my life, give it to AI, and then start collaborating almost with this computer version of myself,” Pritts said. “That has been one of the most liberating things in my practice as a creative, honestly, in my life so far.”
Now, Pritts works most often in video-based art, training DALL-E and other AI software to generate artworks from data sets containing her own visual art portfolio, selfies, and diary entries in the form of thousands of Google Docs pages.
To create her self-portrait “Bitter Recursion,” which is currently on view at bitforms, Pritts followed a recursive process, generating an image with DALL-E, editing it manually, then running it back through DALL-E — repeating those steps until she felt satisfied with the result.
Showing her work at bitforms, Pritts said, has been her first experience dealing with public backlash at such a large scale. To her frustration, much of the coverage she has seen on the exhibition has revolved around the “age-old question:” “Is this art?”
“I’m pretty burnt out on the discourse of whether or not what I’m doing every day is art because it is,” Pritts said. “It’s just such a tired narrative.”
Sacks believes the narrative is shifting, at least in the fine arts sphere.
The most prominent artist showcased in “Artificial Imagination,” Refik Anadol, just made his major museum debut at MoMA with “Unsupervised,” a data-driven digital installation that feeds the MoMA’s collection of over 180,000 artworks into an AI that returns ever-changing generative images to a 24 × 24 foot screen. He also programmed the AI to produce an NFT collection of algorithmic data paintings, the first batch of which sold out in one second.
“Being in a lobby of one of the most important museums on the planet with a pure AI piece is saying a lot,” Sacks said. “I think that museums are getting more and more excited about not just AI, but generative work, media art, in general. They’re finally seeing that one, there’s a value to it, and [two], there’s also a real desire for the public to experience this at a high level.”
Alex Reben, who is showing the second-most expensive piece in “Artificial Imagination” — listed behind Anadol’s $70,000 video installation at just shy of $50,000 — believes the press and entertainment media created a Terminator-esque villain persona for AI that has convinced the public of its fate to take over the world.
“You have to get past that level of the sci-fi of what people understand it to be, and be like, ‘No, it’s not quite that. Here’s what it is. Here’s where it’s quite powerful. Here’s where it’s quite stupid,’” Reben said.
As an MIT-trained roboticist who studied social robots for years, he thinks about his AI workflow as a “relationship that’s a bit of this back and forth.” Reben uses GPT-3, a text generator from OpenAI, to produce wall labels that describe imaginary artworks. Then he creates them.
Reben sympathizes with people’s reservations about tools like DALL-E but simultaneously hopes to combat them with humorous artworks that invite people to engage and learn something. This way, the uncanny becomes familiar, and therefore, less scary. After all, he said, we’ve been scared of new technologies before.
“Progress marches forward, people are resistant to it, and that’s a story we’ve seen throughout history,” Reben said.
“If you go far back enough, almost all art is tech in some way. Every art form where you’re creating something, you’re using a piece of technology, which at some point was new.”
Sacks hopes “Artificial Imagination” shows visitors that AI is a creative partner to the artist, and not something to be afraid of.
“There’s a lot of excitement; there’s a lot of questions; there’s a lot of controversy. But there’s no question in my mind that AI is a tool that artists should be able to tap into.”
About the images
Images and videos taken by Malia Mendez / Peninsula Press.