A show of hands for AI-generated art?
So I’m not sure how I feel about AI image generators like DALL-E and Midjourney.
On the one hand, they offer fabulous art at the click of a button, a well-crafted prompt. On the other hand, where does this stuff really come from? What multi-tentacled algorithmic mega-brain comes up with it? And who is the “artist”?
This is a question I posed to Luis Buenaventura, who is the Philippines’ highest-selling crypto-native artist (his crypto series was part of Curio Cards, which recently sold at Christie’s for a jaw-dropping $1.2 million), and who will be among the many artists participating in the Galeria Paloma exhibit “p a r a / / e l” at Art Fair Philippines this Feb. 17-19 at The Link, Makati.
Paloma’s Lauchengco sisters—Mia, Mimi and Georgia—explained that “p a r a / / el” is centered around “emerging parallelisms in physical and digital/new media art,” and will be simultaneously mirrored at Ninfa Labs, a new media gallery in Milan, Italy.
Paloma’s Art Fair show will feature NFTs, “phygital” and generative works (beautifully displayed on Samsung’s The Frame LCD screens), augmented reality (AR) works that are visible with cellphones and QR codes, and even AI-generated art, which Kimi Lauchengco assured us at a One World Kitchen lunch event entails “more than just text prompts.”
For Singapore-based Wyn-Lyn Tan, this means using specially programmed artificial intelligence programs to “train a curated set of thousands of images of her work” which are then layered (and filmed) to create a generative piece.
Randomness is part of the process; but intention is as well. On a different track, commercial artist AJ Dimarucot plays with Leonardo’s “The Last Supper,” using layered AI-generated art to humorously reflect the Pinoy fascination with fastfood and communal dining.
For Art Fair, Buenaventura developed his piece “Handmade” to shed light on AI’s curious “inability to draw hands.” (It’s true: AI definitely has problems conjuring up convincing hands or feet out of the ether.)
I asked him about the ethics of AI-generated art, and how it interacts with what he does as an artist.
“At first, I felt a little insecure,” he says. “Like, this is the kind of thing that anyone can produce now by typing in a few words, and up until then, I wasn’t even sure I could make something as good as that, right?” But that wore off the closer he looked at the results. His piece is a very funny elaboration of the AI hand-drawing gap: it’s his own created work, depicting a nude female figure within the grasp of what seems like hundreds of poorly-drawn human hands.
“I started looking a little bit more closely at the work that was being produced,” he says, “and now my take is that the ‘A’ in AI doesn’t actually mean ‘artificially,’ it means ‘average,’ like ‘average intelligence,’ because it’s not that good. It’s about 90 percent of the way there, maybe, and the particular work I created for this group exhibit is kind of encapsulating my own anxiety about AI. I mean, it can look great in certain respects, but… it really sucks at drawing hands!” And other things.
You will notice, if you try Midjourney or Motionleap or DALL-E, that faces and body parts seem distorted at times, like they were snatched out of some collective AI consciousness in a rush. Which is what happens when you simply type in a prompt and expect a “perfect” artwork within seconds.
Let’s say you type in “giant red bunny on rollerskates” or “Queen Elizabeth on skateboard” or “the Beatles as cereal boxes” or “devil eating ice cream”—you will get endless iterations, but most of them will be… a little off.
It feels like AI is some sleeping algorithmic brain, just slamming together random bits from its own dark recesses, like dream matter. It doesn’t necessarily make sense, but it’s visually striking. Makes you think about what AI is thinking…
Last year, an artist named Jason M. Allen won first prize at the Colorado State Fair annual art competition for his painting “Théâtre D’opéra Spatial” (a mere $300 was his prize). The painting, submitted under the name “Jason M. Allen via Midjourney,” was created using the AI generator.
Predictably, fellow artists and critics cried foul, calling it “cheating.” But Allen was unrepentant. He recalls conjuring up thousands of images, subtly altering the text prompt to hone his “masterpiece”: “I couldn’t believe what I was seeing,” he said
“I felt like it was demonically inspired—like some otherworldly force was involved.” He shrugs and says he “didn’t break any rules.”
Aside from the ethics of submitting AI art as one’s own, based on the artfulness of your text prompt, Filipino artist Buenaventura thinks it’s only a matter of time before AI gets it right. Those hands won’t be badly drawn forever.
“This is encapsulating a moment in time,” he says of his “handmade” piece. “Five years from now, maybe five months from now, AI is going to figure out how to draw better.”
But, he assures us, they will never be able to do it without humans: “The AI consciousness itself cannot come up with a message that someone is trying to convey because it is not conscious, there’s no actual consciousness. Thank God. Because if it did, we’ve crossed that line into Terminator reality.
Maybe. I think of that guy at the Colorado State Fair who thought it was groundbreaking to win an art contest, even for $300, by pressing a button, letting a computer intelligence do all the rest. Part of me thought the guy shouldn’t get to keep the money, or somehow should share the money with the AI.
But AI doesn’t need money; it just needs to do what it’s doing, right now, better and better.
Which is, ultimately, to replace us.
* * *
Galeria Paloma’s “p a r a / / e l” will be showcasing works at Art Fair 2023 by the following artists: Skye Nicolas, Luis Buenaventura, Bjorn Calleja, Jopet Arias, The Alarcon Brothers, Carlos, Wyn-Lyn Tan, Raymond Lauchengco, AJ Dimarucot, Aswang, Sevi Agregado, Isaiah Cacnio, and Sheila Ledesma. For more information, visit www.galeriapaloma.com.