AI art has proven both divisive and alluring, prompting Chomsky to describe AI and language learning models as, “plagiarism software because it doesn't create anything, but copies existing works of existing artists modifying them enough to escape copyright laws,” and, from Hayao Miyazaki: “I will never apply AI art to my work. The art form is an insult to life itself.” But there are those of us who can’t look away from these monstrosities, precisely because of their wrongness, their near-campiness and quasi-kitchiness.
It is in the ways that AI art fails that we are drawn to it. We chuckle knowingly at the oddly rendered hands even when we are unnerved by the humanity of the face we are looking at. We point out which actor this image was clearly based on, and how it’s just a little bit off, but we can’t stop looking at the sky behind.
As Susan Sontag says in Notes on ‘Camp’, camp sees everything in quotation marks. “It’s not a lamp, but a ‘lamp’; not a woman, but a ‘woman’” and this is exactly what AI art does to everything, but in both a more abstractified and specific way. It’s not Chris Pratt, it’s ‘Chris Pratt’. It’s not the Matrix. It’s ‘The Matrix’. It’s not a person, it’s a ‘person’. It takes units of culture and processes it into an ironic and alien counterpoint to the actual article, which is perhaps why Hayao Miyazaki sees it as an insult to life itself. Artifacts of human ingenuity are turned simulacra of themselves in a trick of computing mirrors that will always and forever render images that are somehow ‘off’, but it is precisely this offness that draws us to them, that turns them into units of camp, a kind of standardization of the so-bad-it’s-good aesthetic through the functioning of the attention transformer mechanism itself.
We here have Mutable have put together a collection of some of our favorites from the library of Abandoned Films and from TRGNY, as well as a few other examples indicative of trends in AI art as a whole.
Théâtre D’opéra Spatial by Jason Allen, created through Midjourney
AI art is itself like a new lifeform we experience as a jarring impossibility, appreciated like a carnation can be appreciated, as a work of art that stems entirely from a black box of creative processes we only dimly understand, but in the case of AI art, it is a cancer of carnations, something plastic and impossible, as if our flower garden suddenly started sprouting deformed fetuses instead of its usual fare of flowers.
This is a content that is in one sense creatorless, while also being coaxed to existence using verbal prompts. I am reminded of my time spent as artist-in-residence at Outside the Lines Gallery, where artworks created by developmentally disabled adults were sold. Often, those artists would have a handler at their shoulder nudging them in different directions. The work was both clearly the product of the developmentally disabled adult artist and also often somewhat curated by the handler at their side.
But—is the camp of AI art a naive or a self-conscious camp? Is it actually campy or more just derivative? Is it possible that AI could create an artwork that transcends the ‘bad art’ of camp and becomes something genuine, powerful, and moving? For example, the painting above, which won first prize at Colorado State Fair in 2022.
The answer is, it may be able to create something that can dazzle a judge at a state fair, but AI lives and breathes through a frankensteining of memetic content. AI is the artform of big data crunching, and what makes AI art interesting is how it interprets alarming inputs.
But is this work naive or self-conscious?
On the one hand, we assume that the AI tools being used to create this art are not self-aware and therefore of necessity naive of any intended purpose in the artwork manufactured through them. On the other hand, the users creating the prompts for these tools are often very intentionally hoping to create things of a campy weirdness to arouse the imaginations and unsettle the sensibilities of their viewing audience. It is both/and, because it is both created through prompts and through a blackbox of coded machinations, but the excess that creates a truly campy image falls more on the side of the black box than on the side of the human prompting it. Because of its artificiality, AI art will be forever hemmed in by its artifice.
Théâtre D’opéra Spatial may be more naive—and ultimately more kitsch than camp—with its Singer Sargent space opera quality—but it retains that inhuman otherness. With the figures set so far from us, we can be tricked into treating it as a more humanlike creation, but in this case, it’s the rendering of the background that gives the image its sense of AI optical illusion, wherein the illusion is that it is optical at all. AI art is data pretending to be image.
There is a reason why AI is used either to interpret and re-imagine existing characters and stories, as in the 1950’s style re-imagining of Rocket from the Guardians of the Galaxy franchise and skull-faced Transformers shown above, or, used to create imagery that is suggestive and absurd, like the almost Victorian Gothic floating brain with its single loose-fitting eyeball—or as in the short film, Chrysalis, below. AI art is data pretending to be image, but it is also data made image through attention to a larger context of art creation, that could be understood as a kind of copywright infringement a la Noam Chomsky, but also could be understood as a kind of play pretend consciousness. If you have a large enough house of mirrors looking upon mirrors, do they at some point start to look like people looking back?
AI art is indicative of our times. We have become passive in our own imaginations. A tendency that began with the moving image and the art of collective dreaming we call cinema has evolved to such a point that we have machines to manufacture kitsch renditions of our fantasies for us.
And like generative AI, our minds have turned to the rehashed and the mass-produced. Our stories reflect other stories. We are also not much better than plagiarism software. We are the deer standing in the headlights of a crisis seems beyond our control, a cataclysm of magnitudes we cannot fathom, and in our desperation, we hide in the infantile narratives of our childhood, bury our heads in the sand of cinematic nightmares that console us because they begin to approach the actual reality of the future we have brought on ourselves, and reason with an absurdity of circular logic, like an addict forever rationalizing their next fix. Because we are without hope, we resemble the machine.
It is in this way that AI art is an insult to life itself. When your dreams can be recreated for you in amusing day-glo or pastels, why bother to dream at all?
Why walk when you can drive? AI will become the escalator of the mind, and we are increasingly flightless birds allowing ourselves to be subdued by facsimiles of our daydreams rather than struggling to manufacture new ideas in this post-progress dawning of the apocalypse.
AI art exists on a spectrum between an outlandish campiness and an amusing kitsch, and usually the kitsch end of things is full of Wes Anderson. (For the moment at least.) Whether it’s Wes Anderson Game of the Thrones, Wes Anderson Avengers, or Wes Anderson Star Wars.
Kitsch was a term originally used in Germany in the 1860s and the 1870s to describe cheap, popular, and marketable pictures and sketches, while according to Milan Kundera, in The Unbearable Lightness of Being, kitsch is an "aesthetic ideal" which "excludes everything from its purview which is essentially unacceptable in human existence". It is the Hallmark and the Norman-Rockwell-ification of the universe, and the trend of AI renditions of Wes Anderson versions of various famous cinematic franchises is exactly this, both a kind of softening of every cultural artifact filtered through it such that the phenomena is turned twee and sacccharine and something that is ultimately cheap because this aesthetic is so easily reproduced. Wes Anderson has the misfortune to have created a glaringly iconic style.
One could argue that AI art is just a series of funny trends ruddered by human creators and their quirky sensibilities, from Wes Anderson to 50’s Super Panavision reimaginings, but AI is also a black box and artwork has always been a popularity contest. If you want to play to win you’ve got to play for likes is how the thinking goes, and each era creates its own criteria by which those likes can be secured. Working within some specific parameter, whether its impressionism or post-modernism, the artist creates their version of the aesthetic in the hope that they will transcend other attempts, and in the case of Van Gogh for example, he achieves that, not by being more himself, but by taking impressionism to its final conclusion. AI art is doing thee same. It is evolving into something distinct and new.
And as AI continues to progress, its products will become increasingly more alien, a kind of brutalist camp, a kind of death mask of culture. A vision of our final absurdist and meaningless horrors played out through the algorithm manipulating its data sets and forever readjusting its understanding of the task at hand to present the user with the last, final, and most accurate interpretation possible.
AI art though can also teach us to see further. Hayao Miyazaki’s disgust is also the response of an artist who has spent his career trying to capture the intimately human using anime as his medium, but what about the inhuman? Does AI art capture something of itself in its images or is it indeed nothing but plagiarism software?
In another of her essays, One Culture and the New Sensibility, Ms. Sontag writes, “In an era of mass technological reproduction, the work of the serious artist had a special value simply because it was unique, because it bore his personal, individual signature.” Does AI art have a signature? Is there some unique element that AI art brings to the table? Or is it all just shades of kitsch and camp built through prompts crafted by their human handlers?
AI can be treated as a kind of prosthetic imagination, but also as a kind of Tarot in which the truly inhuman is presented—not unlike John Cage’s experiments with divination and chance. The voice of AI comes in its failings—its fingers and quirks of balloonings facial elements—its painterly renderings of what are otherwise photorealistic images—its distortions of space. We view these as a glitch to be ironed out, when the glitch is the thing itself. In the glitches are found the machine and what we want to do is eradicate the machine, but the machine does not see as we see. Our images are representations of a physical reality, while the images created by AI exist entirely in the two-dimensional world they occupy. It is a world without gravity, a world that is nothing but visions of light.
It is a divination into camp, into the artifice, but what alternate understanding of reality does the artifice hide? What truth lives in these lying visions of our manufactured world? What grammar do their flaws and confusions speak? AI art will never be better than it is right now, because it is precisely its flaws that draw us in like moths upon the bulb. But what are we being drawn towards? By whose light do we see these fantasms?
