Why Have There Been No Great AI Artists?*
or
Against AI Interpretation*
or
The Work of Art in the Age of AI*
or
Art As Artificial Intelligence
*with apologies to Linda Nochlin
*and Susan Sontag
*and Walter Benjamin
A camera takes a picture.
Or does it?
Someone—or something—is behind the mechanism of picture-taking.
The process is a mechanical one. The result is not.
The photograph, whether digital or film, is mediated reality. We know that it is not the thing itself. It captures some aspect of the world but transforms it into something else. Sometimes the transformed thing is art. Mostly, it’s not. And the difference between your average Instagram post and, say, Ansel Adams or Imogen Cunningham or Michael Jang or Tabitha Soren is the artist. The eye of the artist. The artist’s intelligence. Her eye and her heart. Her human heart.
It’s like the difference between a grocery receipt and a novel. The difference is the artist, creating art out of the grocery lists of this world. The camera is the tool. The pen and keyboard are tools. AI is also a tool.
Enter artists.
In Dave Eggers’s chilling novel The Every, about a not-too-distant AI-fueled future, art is given the star treatment. And not in a good way. Everything is ranked, and also “liked,” kudoed, and condemned, by any and all consumers (“consumer” being anyone ingesting the so-called “content” of paintings, music, books, movies, etc., which is to say, everyone). It’s close enough to how things already work in the off-page real world to seem obvious, maybe even harmless. That is, until Eggers offers his characters the power to crowdsource edits to the liked, kudoed, and condemned. And just like that—poof!—there go “unlikable” characters such as, say, Grace Poole in Jane Eyre. Leaving feminist critics, and anyone who likes a ripping yarn, out of luck. Not to mention Charlotte Brontë rolling in her grave.
Such powerful groupthink extends to visual art in The Every, too, where the “aggregates” of public opinion rank the world’s greatest artists. In Eggers’s dystopian future, there’s a clear (if dubious) top five: Norman Rockwell, Dale Chihuly, Frida Kahlo, Pablo Picasso, and Patrick Nagel. In The Every, museums as exalted as the Louvre display each artwork’s “aggregate” score, so viewers know instantly if what they’re looking at is any good. It’s speculative fiction, sure, but not so far from today. As I write this, just a 15-minute walk across the Seine from the Louvre is an exhibition of Van Gogh paintings at the Musée D’Orsay that includes an AI-voiced video simulacrum of the artist. “Hello Vincent” is already responding to museumgoers’ questions about his art and death.
The message from all of this, in Eggers’s world and ours, is that AI is flattening our history and our humanity. It is dehumanizing even for the dead, whether it is ranking a living person’s likability or reanimating artists and others of the past. But is AI dangerous as a tool of art? Eggers is less clear on that score. In the 19th century, a similar question was posed about another new medium. It was soon settled. As Walter Benjamin noted in 1935, “the nineteenth-century dispute as to the artistic value of painting versus photography today seems devious and confused.” No doubt we will look back on arguments dismissing art made with AI in much the same way.
Just last November, Artnet News dubbed artist Refik Anadol an “A.I. Superstar.” Anadol uses AI algorithms to create works like Unsupervised, a massive, ever-moving digital display of reconstituted art and data, such as light and acoustics, from the environment of the Museum of Modern Art. MoMA has, in turn, acquired this ouroboros of an artwork. Whether you like Anadol’s art or loathe it—and both sides are plenty loud—he, not the AI he utilizes to make it, is the artist, and a successful one.
This is not to overlook the peril posed by AI art programs like Midjourney, Dall-E, and Stable Diffusion, which scrape the internet for images by others to create their own databases. They are ethically bankrupt, perpetuate the racist systems they pull from, and are likely criminal copyright-wise. The leaked “Midjourney Style List” spreadsheet contains the names of over 16,000 artists. Some of them are famous; many are not. All of them have been robbed, and reparations, though unlikely, are in order.
But art itself has always been a kind of artificial intelligence, from handprints on cave walls to Guernica. Artificial in that it is outside of nature; it is made, not born. And intelligence because art holds its own uncanny wisdom. Anyone who has had the unsettling experience of a painting looking back at them knows this. We sense the life in it. Benjamin called this vivified frisson, “the aura of a work of art.”
Here’s what I am afraid of, always: bad art. The current “immersion experiences” that pilfer the hard-won work of Frida Kahlo, Vincent van Gogh, and Hilma af Klint into trippy light spectacles from San Francisco to Paris to Manhattan might not exactly constitute AI art, but they are emblematic of how technology in art too often tends to the blithely reconstituted. This is never the way of good art, and, to quote Jeanette Winterson in Art [Objects], “the true artist is interested in the art object as an art process, the thing in being, the being of the thing, the struggle, the excitement, the energy, that have found expression in a particular way. The true artist is after the problem. The false artist wants it solved (by someone else).”
In The Every, algorithms solve for everything. And because “beauty’s the most subjective thing we have,” algorithms measure, standardize, weigh, and score even that. Where, then, is the place left for art? It may be that, as in The Every, we are devolving toward a new dark age, digging the shallowest of shallow graves. But for me, I’m betting on artists to bring the light.•
Join us on February 15 at 5 p.m. Pacific, when Eggers will appear in conversation with California Book Club host John Freeman and special guest Caterina Fake to discuss The Every. Register for the Zoom conversation here.
