Why Is AI Art So Revolting?
Aura’s afterimage
With a clickbait title like this I think it’s only fair to reveal my position as early as possible. This article isn’t a call for an end to AI generators or models. Nor do I think AI can never be art. Importantly, I believe that it is not currently capable of being art. Or it only renders unsuccessful art. I believe that revolting is the correct word (and it is the adverb ‘so’ which makes the title clickbait) because I believe that the layperson who interacts with AI art almost essentially does so in ignorance or against their will. We are crossing a threshold where generative AI models are passing out of the exclusive domain of nerds and into popular consciousness. What will prevent a flood of mediocrity now that everybody can become an artist? It will be the same which has prevented everybody from becoming a photographer even with each new development in camera technology. One day anybody will be able to become an artist via the technology of generative AI- the playing field is being leveled in the same ways that it has by photography, commercial pigment, print, ink: the tools are available to all, the capacity to make art with it is limited. Until practitioners of AI art face this reality, they will continue to produce facsimiles of real art. Works which, while echoing at the aesthetic tendencies of real art, flounder: repulse.
Hands
AI models have crossed the final threshold and can now convincingly pass as photographs. Many now boast the correct amount of hands, and when they are noticeably AI, it is often difficult to tell whether it is a failing of the generative model or of the operator’s taste. So then why does the discovery that an image we enjoy was made using AI feel like a betrayal?
The utopic selling point of AI is that an artistically minded person may conjure images directly from their imagination, and if that is the aim then current models fail. As a photographer, it may seem hypocritical to say that AI models create categorically bad art when photography as a medium faced the same charge for the first century of its existence. How can we be so short sighted as to pull the ladder up when the arguments are already settled? Just as photography does not simply photocopy the world but instead creates an image that is every bit as manipulatable and as distinct from reality as a painting or a poem, so too might AI prove itself just another tool at the artist’s disposal? One day, and maybe even soon. But for now, all an AI artwork can successfully do is trick us into thinking it is a different, more honestly rendered work of art.
While they are problematic, I do not intend to engage with the more common arguments against modern AI models: That they train on stolen art, and that they draw from a limited pool (ie. only what a model has been trained with, albeit in an unrecognisable form, can be produced). Even if we started fresh with models which were ethically trained with artworks from consenting, compensated artists, AI would still fail to create successful, compelling art. Even if AI models were able to create wholly original works, they would still disappoint. This is because of who is advocating for ai art in the first place.
A Token of the Real
Right now the biggest fans of AI are the crypto crowd, tech bros, and hedge funds (to ignore the significant overlap therein). These capital obsessed personalities are interested in art only in the ways the art market and its exploits might represent an increase in their personal wealth. Procedurally generated nfts are amongst the most artless images on the internet- somewhere below advertising and porn: both of which at least attempt to titillate before making a swipe at your wallet*
“They have tried to replace writers with executives, who know less than AI, believe me. So any way they can replace writers, they would be happy to do so.
I don’t have a great understanding of AI, I have to tell you. I mean, I understand it more than I did initially. But as far as I can tell, it’s just stealing. They take a lot of things that writers wrote and then steal them and then mix them up and then put them together in some other way so that it seems like it’s something else.
Employers would like to replace every employee if they could. That’s the biggest expense to any business, no matter how poorly they pay the employees. That is the trajectory of capitalism, you know: How can we get the most money with the most profit? That is usually by having the fewest employees. They never think, “Why don’t we replace the executives?” — Fran Leibovitz, Professional raconteur Fran Lebowitz thinks art should be useless, Vox
AI is championed by people for whom creativity is not a source of nourishment but a liability. It is employed by those who look on at the personal and fiscal benefits of creativity with envy and scorn. It is little surprise that AI has taken off in spam, scams, and as a lazy bypass to employing creatives.
Elsewhere, AI art has seen the liberatory promises of NFTs eat their own tail. By flooding the marketplace, the already flimsy proposition that NFTs could benefit artists has given way to accelerationist capitalists whose only aim is to exploit the system to benefit themselves as efficiently as possible.
Find me a real artist experimenting with AI rather than a hypercapitalist happy to fit anything into their grindset, and perhaps we can continue the conversation as to whether AI can ever make art.
*are banner ads which link to malware our closest aesthetic equivalent to both the artfulness and function of procedurally generated nfts?
Real artists experimenting with AI
My favourite wedding photographer Emin Kuliyev has begun experimenting with ai with occasionally stunning results. I’ll freely admit that I enjoy some of this imagery, and Emin’s artistic eye is intriguing. These images even in many ways match the tonality of his photography. However, I have still found myself repulsed by these images upon discovering that they were generated (either through his caption or through closer inspection). His images, while fantastic, have always worn their surreality on their sleeves: The joy that I, as a photographer, take from his images is in decoding the intricate way they were crafted. These images individually are not necessarily outside of my capacity as a photographer- but Emin pulls them off with such consistency and always with such enduring creativity that I find his work utterly thrilling. This same sense of thrill is whisked away by the discovery that an image was created with ai- the poses, the boutique optics, the precise timing and the perfect juxtapositions of real life- all are, as one commenter said, “available to you for $24 a month.”
“Even the most perfect reproduction of a work of art is lacking in one element: its presence in time and space, its unique existence at the place where it happens to be. This unique existence of the work of art determined the history to which it was subject throughout the time of its existence. This includes the changes which it may have suffered in physical condition over the years as well as the various changes in its ownership. The traces of the first can be revealed only by chemical or physical analyzes which it is impossible to perform on a reproduction; changes of ownership are subject to a tradition which must be traced from the situation of the original.” — Walter Benjamin, The Work of Art in the Age of Mechanical Reproduction
When Walter Benjamin wrote about the death of an artwork’s aura he was talking about photography, with its infinite reproducibility and lack of a singular, original copy existing in its own immutable time and space. It can again feel like pulling the ladder up behind me, for a photographer to scrutinise a new medium using arguments which were first turned on photography. However what artists who employ ai will have to prove in many ways mirrors what photographers had to prove in order to justify their own entrance into the art world. What is the art value of a photograph beyond its literal recording of the world? In the same vein, what has to be imbued into an ai prompt for it to feel like the prompter has created something?
Democratising Art
In an essay on Atget republished in Core Curriculum, Tod Papageorge relitigates an age-old argument against photography’s place in the art world as posited by Rosalind Krauss: arguing that argument over whether photography merely indexes a space has only arrived as the technology has matured, he says “photography’s status as an art only turned into something resembling pitched battle when what might be called the full conditions for the art were finally present, in other words, when the act of photographing more nearly resembled the human act of responsive, continuous perception than the static one of transcription more or less imposed by earlier technologies.”
This technological progress is echoed in the development of ai models. We have seen a tremendous leap from early DALL-E models which rendered an impressionistic dream-like vision only particularly useful for memes, through to modern models which can create photorealistic imagery. AI Art now resembles real art.
“As the photographic industry was the refuge of every would-be painter, every painter too ill-endowed or too lazy to complete his studies, this universal infatuation bore not only the mark of a blindness, and imbecility, but had also the air of a vengeance … I am convinced that the ill-applied developments of photography, like all other purely material developments of progress, have contributed much to the impoverishment of the French artistic genius. The only place for photography is to be the servant of sciences and arts — but the very humble servant, like printing or shorthand, which have neither created nor supplemented literature.” Baudelaire 1858 (As quoted by Bill Jay in Occam’s Razor)
Baudelaire’s disgust in photographers translates well to the current criticisms around AI art, which bypasses creatives who have spent the time learning their medium. This is more than jealousy and fear of obsolescence from artists, it beckons, why does AI ‘art’ want to be considered art in the first place?
“This is where these projects diverge significantly from the artworks that drove the speculative mania in the spring. They’re not positioning themselves as art that is worth possessing for the art, as flimsy and illusory as that claim was, but as ongoing projects that you buy into. The purchase of a token is an abstracted version of buying early stock in a company, a venture-capital style investment in promises, which is a very significant mutation.” — Folding Ideas, Line Goes Up [Video Transcripton]
AI is supposed to democratise art, but art can be made with a pen and paper, with ones own body, with mud on a cave wall. The sales pitch of democratising art only appeals to those who have felt shut out by art institutions and who confuse- through lack of contact- the huge sums attracted by public art sales. In reality, the way that photography and AI deviate is that photography’s tenuous position in the art world was advocated for by photographers. If all photography is taken out of galleries tomorrow it will not neutralise the work of thousands of photographers over hundreds of years. The push for AI is to legitimise a massive venture capital spend in the hopes that it will reward its backers at the total expense of working artists.
The world is full of artists who make art because they are trying to make sense of the world using whatever skillset they have developed. They rarely become fabulously wealthy. They often do not even receive recognition relative to the effort put into their work. Yet they persist because art is the mode through which we carve out a little bit of human experience and try to understand it. Until ai art is used to this end, rather than to replace real artists for their skillsets, rather than to try to get in on the infinitesimally small ‘action’, this is the disconnect which makes AI art look but not feel like real art.
