Life

Why Is AI Art So Cringe?

Un collage d'un vieil ordinateur peignant un vilain tableau sur un chevalet avec une palette de peintre - sur un fond bleu dégradé.

When AI first emerged as a practical possibility on the scientific scene in the late 1950s, it was hard for the excited general public not to conjure up images of a near future of flying, self-driven cars or robots that would save the world. Now it’s 2023 and, unfortunately, instead of solving the greatest issues of our time (like, oh, reversing climate change), we’re stuck doom-scrolling through countless “AI selfies”, made using the Lensa app. Sure, there’s an argument to be made for Siri and Alexa — or even AI-powered facial recognition that unlocks your phone — being helpful additions to modern life, but the latest buzz around AI-generated art is happening despite its non-existent purpose. Perhaps more importantly, it’s also completely soulless. So why are the vibes so completely off with AI art?

Since the recent launch of image-generator tools like DALL-E 2, Stable Diffusion and Midjourney, the internet has done what it does best — used them as cursed meme machines. They’ve also sparked larger debates about what constitutes as art and around plagiarism, as OpenAI was trained using the images of copyrighted artworks. With Cosmo releasing the first (frankly forgettable) AI-generated magazine cover last year, these creations have yet to create anything remotely compelling. So far, the only thing it’s been really good at is pissing people off: Lensa’s software was criticised for hypersexualising its user portraits and artist Jason Allen faced backlash for winning Colorado State Fair’s digital art prize with AI-generated work. 

Videos by VICE

Allen compares using AI art technology to channeling an “otherworldly force” or a demon. With much of the work looking soulless, cursed or possessed, he might be onto something. Lolita Cros, an art curator and advisor based in New York, says that despite this, AI art was always bound to happen. “Impressionism is a product of the tube of paint and digital art is a product of the internet and computers,” she says. “Not only will art be forever informed by the medium, but the medium itself can allow for an entirely new art movement.” Cros predicts the AI medium will eventually become accepted as a “common tool” – yes, really.

While Cros is convinced that AI art is here to stay, she’s yet to see anything “good” yet. “I think it’s because we’re in the early stages of it and the medium is not yet the means to an end, but the end itself,” she says. “Once people master the tool I’m sure some interesting art will come of it.” Cros compares our current obsession with AI-generated selfies with work such as The Portrait of Cosimo I de’ Medici by the artist known as Bronzino. “Commissioned portraits in general are cringe and I’m yet to meet an artist who enjoys doing them,” she says. “People don’t commission portraits to showcase their bad looks, they do it because they want to look hotter, stronger, taller. If Bronzino had painted the Medici family with spinach between their teeth I don’t think he would’ve had much of a career.”

If we can imagine a painter in the 1500s groaning about having to create an effectively airbrushed portrait of someone royal and prestigious, then it begs the question; is AI art so lifeless because we’re such dreadful, unoriginal and demanding commissioners? Meaning that the current soulless state of AI art is the fault of the people behind the algorithms, too. “I think what makes art good isn’t that a human physically made it, but rather the idea that made it happen,” says Cros. “If algorithms can be assembled by a thoughtful human with good ideas, then I can’t wait to see the AI-generated work they’ll come up with. The problem is when the humans creating those algorithms are uninteresting and uninterested in art.”

Paul Hill, founder of Gen Z art gallery Strada, says there are so many pieces of bad AI artwork right now because there’s always been an abundance of terrible physical artwork. In other words: Whether you’re handing a paintbrush to a crowd or an image generation service, most of the results won’t be gallery-worthy. Hill prefers to think of the tools as an “internal resource”, instead. “The biggest benefit is using it as a tool to flesh out ideas and variations before fully leaning into something, like a physical object, all the way,” he says. “I think removing people from the process of making artwork or clothing, and just solely relying on a machine, takes the art out of that creation.” 

Hill views AI image-generation tools as an opportunity to “reduce time wasted by the humans involved”, not as a way to remove humans from the artistic process altogether. This is an approach that could eventually save us from the mediocre-ness that is currently being inflicted upon us by “cutting-edge” technology. After all, it’s often the stories behind the art, or the artist, that draws crowds outside of galleries and museums. Think of the alluring “tortured artist” narrative seen in the dialogue around Vincent van Gogh and his infamous ear or Banksy’s start as a mysterious, anonymous artist.

Throughout history, it’s been proven time and time again that art becomes more compelling (and profitable) with the right narrative. The reason AI art is so terrible right now is because it’s not being used as a tool to showcase a message – instead, it is the sole narrative. Let’s face it, a bunch of tech bros coding software that internet users have turned into a meme factory makes for weak storytelling. So, too, does the rise of brands using AI art as an edgy way to sell products. With this in mind, AI art will only become an innovative artistic tool once it’s actually used for artistic innovation. Until then, we’re merely children creating stick drawings to be hung on our fridge (AKA Instagram).

We may have to face the fact that the current state of AI art is cringe because we’re cringe. Really, we’re no better than the unimaginative and self-obsessed people from the history books that commissioned bad portraiture. This first phase of AI art may go down in history as another lesson, too – that just because something is possible through technology, doesn’t mean it should be actually made. As Hill puts it, “there’s so few good digital artworks because there are so few good physical artworks”. Basically, there’s a load of crap physical art in the world, so now we have to suck it up and witness a load of crap digital, too – what a time to be alive / time to throw our phones into the ocean.

@lauraepitcher