Against AI Art - Part I
An argument against AI art, & a compulsive defense of every artist and artistic copyright involved in the development, and output of, AI art programs. With other thoughts on capitalism & this moment.
It’s 100% Derivative, 0% Ethical
Before I dive in, let me start by saying that I’ve put a lot of hours into learning and making both visual and audio art. It’s safe to say it’s consumed the majority of my life and been a top priority and compulsion in it, by a wide margin. This Part 1 piece of what I hope will emerge - after Part 2 - to be a cohesive and well-made argument is very much shooting from the hip to get the ball rolling…winds up more reading like jumping into a fire. While it’s true that I have seen a lot of society’s unwashed underbelly, I believe mostly that I err on the side of realism over cynicism. And then there are the lawsuits against Stable Diffusion, MidJourney, and DeviantArt for copyright infringement, wood and oxygen to the fire I’m jumping in. In Part 2, though, I want to explore more the philosophers and social / psychological / cultural thinkers that have influenced my thoughts on, and obsessions with, art-making as an inherently humanist effort. Certainly I can see a present and a future where AI merely assists with art-making, and does not inhibit any singular artist’s development through misapplication. However, I remain a skeptic of ‘net-positive’ value judgements on AI’s future-impact on our human society as it is today.
So let’s get to the meat and potatoes.
And decisively.
Ai art is entirely derivative and lacks a fundamental component of art: creativity, of which much has been said and determined over the millennia of humanity’s relatively young existence (in aesthetics, philosophy, anthropology, art history, psychology, neurobiology, etc.) The question of whether something is creative has taken the world’s great philosophers across terrain spanning the rational to the imaginative (which of the two faculties accounts for its allure and expression, creativity?), from the ethical to the spiritual (is creativity a kind of virtue? A happy accident of the human’s aim to cope with conditions and circumstances beyond his/her/their control? An attempt at communion with reality and subjectivity, a kind of synergy of both?) One could parenthetically address and explore the question of what creativity IS at length and say absolutely nothing new, because it’s been studied for centuries across disciplines almost to death. You could start with Plato, and end up deep in the throes of experimental neurobiological studies and still be faced with what essentially is a blank canvas fraught with questions. And is THAT in itself a kind of experience of creativity. IS it whatever an individual decides it is, and no longer subject to a consensus, or even in need of a society in agreement about the definition of art? Or is it a social necessity and an endeavor that need be available to all individuals in the future as a means as well as an end-in-itself? And if that is the case, why not press pause and figure out a way to compesate all users supplying training data? It’s global impact makes it an issue demanding more global thinking.
It has the flavor of gimmick
Maybe that’s why everybody is into the way the artificial intelligence is making art - it’s devoid of its own personal aims, yet does most of the legwork. It’s a-moral, without pretense (and at the same time, without a prevailing-over-conditions, which is that timeless component of art people stand for hours before a painting or in an installation in a museum for - if not because in the artist they recognize something of themselves/want to support someone like them / have admiration for their personal struggle ). W e can say it’s the ‘prompts’ that determine the outcome and to an extent those are a kind of creativity, but the end-goal, the output is determined by the co-opted individuality and artistic expression of other artists or the input. It is the work of other artists, ALL of whom have never consented to have their work copied and ran through an artificial machine for the sake of profit, clout, or some derivative creative endeavor, that is truly creative, generative, and contains that human element of adaptation and non-synthetic sense- experience and perspective which make art the panacea and immortal dialogue of the collective human imagination. Does ai-art successfully beg important questions, and add to that conversation? No doubt generations in the future will have much to say about its origin and genesis, being that in it’s current form it is fundamentally exploitative to the fullest extent, in a way that even institutionally-supported artists who employ un-credited skilled laborers for their works could have only dreamed of being (but expressly, consciously, didn’t).
It lacks in most cases anything resembling the human ‘touch’, making use only of human ability to use language and combine words - exactly the opposite of what makes art, art.
Creativity, from the beginning, has concerned the subjective sense-experiences, perceptual and stylistic idiosyncrasies of the lives of individuals with respect to their relationship to society or their experience of some aspect of their environment, as a changing-yet-stagnant collage of meanings. AI-generated art, however much it pits the creative individual in opposition to society vis-à-vis a lifeless, computationally-randomized, wholly-derivative set of algorithms, merely co-opts the individual and subjects it to society. It is a one-way conversation, because while the ‘art’ of applying words to prompts for improved results could be said to be unique, that is not itself the same as Doing art.
What sets art apart from the mundane and elevates it to the place of its positioning - being a kind of commentary on life and reflection of individual or collective experience - is that it does not make use of words which connote and denote whole shifting categories of meaning that requires cognization, or an interpretation and selection )sometimes mindful, sometimes habituated) of possible meanings. Art can be experienced without words and is intended to be in the case of visual art (that is not film), as a kind of dream-like state of dialogue with another mind, in which meaning is conveyed and expressed without the need of words. While more and more are people engaging with art with words, they do so knowingly when they do. They know with whom they are engaging, and the identity or identities behind the work is known. With a comment here and there upon the work (quite different than words being the sole ‘unique’ tool of the work’s creation other than those works conscripted non-consensually by the program to meet the demands of the prompt), the words with respect to visual art normally are either supplementary or ancillary. And this is what makes art special. They are not, as with language, carrying the weight - badly, limbs akimbo, gait stilted. And they are not doing so alongside an algorithm trained on theft.
Visual art (that can reasonably be called art) translates meaning sensed directly and as conversation between two human beings with fundamentally the same organic senses and capacity for sensation. It is a sort of stripping away of cognized symbols with respect to words and words with respect to symbols (and one need not be an explicitly verbal learner or thinker to understand this); the mechanics of working raw materials - even if it is lines or colors synthetic and piecemeal in a computer software program - is explicitly what makes the process of making art, artistic. It is the great, and vague, subjectivities drawn out by encounters with the raw material arrangements at play. It is bringing the imagined down to the realm of the real.
What could have been, and still could be, an enriching dialogue between the individual that can be described as an ‘artist’ and the world in which said artist is known, is a domain with gates now open to individuals with no compulsion or need to get familiar with the tools and skills of creating something and engaging in the methodical application of motor skills and tools that without the human mind, in themselves, have no specific purpose (not to be confused with function).
This AI Art takes the human mind as its only tool, and on either side of the equation - whether ye be the artist providing the source material or the prompt-writer forging the commands - reduces human intelligence and its primary improvisational function down to the material of some a-moral, degenerate code that randomizes the several billion works of art it has codified and categorized (again, without the consent of these artists) into something only resembling art.. It’s a facsimile, striving towards comparison and succeeding only in almost-competition. But indeed, almost.
Anyone with a mouth can take an easy short-cut into the world of digital art, visual art, and even music without so much as a youtube video on the tools inside Blender or the mechanics of light, composition, color, or an art utensil. If you have the words, you have the will apparently! But not quite. You also need to know OF someone with creativity of the kind associated with art, suddenly you can now create your own derivative works of art that lives and breathes their style and flavor. Pawn it as your own! Just for fun, you think. Just a laugh. Until, you know, some Dr. Evil advertising agency guy gets ahold of it.
And that’s not to suggest that most art isn’t in some way derivative, because most of it definitely is…the common thread between all art that isn’t simply a lazy facsimile of art is that it demanded the development of skill.
The main question though:
Is an algorithm built on the back of other artists that generates profits from, but does not compensate, said artists, really creative, in the hands of someone creative? How can it be, when by its fundamental nature and the manner of its construction, it degrades and abuses the very core of the artist-identity? Does a tornado create a new apartment complex because it destroyed one?
No.
Its just destructive.
Conclusion - No, thanks.
There is no consensual dialogue or mutuality between the artists whose works were used to train the models, whose work is pulled from with the various prompts, and society. In the case of AI art, the art doesn’t say anything new about society or the individual constructing it. It simply regurgitates in some new form the same data it has been given access to - again, without the consensus of artists or societies or formal associations of any kind. Very little in history do we find examples of unpaid skilled laborers being developmentally, emotionally, socially, or spiritually enriched by degrading treatment from their exploiters. I don’t think this will be any different, in the long-run.
The final question: what might capitalism do with this next?
I think most people hoped that AI would be applied more to unburden those with a burden too large. Or maybe that’s not precisely what we had in mind, but what DID we have in mind? Is anyone flabbergasted that this is where and how the ‘innovative’ qualities and potentials of AI would be applied in viral fashionin the dawn of it’s inception in a capitalist society? Probably not. But, to appropriate the works of artists is to TRY and lay claim to the necessary function they have historically served in society - to reflect, often critically, on the present time. Not as only general ‘artists’ but each and every artist as an individual, by name specifically. A society is not judged by the fact of its having any artists, because all societies do, but by the specific artists making waves and causing stir in the ‘culture’, as it were. And that is not only in the sense of their names, but their backstories, their personal struggles, qualities, private and public lives, their memories, (and maybe others?)
Work that is an extension of their imagination and their bodies and their memories, dreams, autonomy (or what remains of it), and judgements? I mentioned earlier that to appropriate the work of artists without regard for their consent is to try to lay claim to the functions and roles that artist’s can, daringly and perhaps more rarely, choose to take up. But more importantly, even if the role of the artist cannot be effectively taken up by projects of Ai-art - as i’m sure will remain the case- the MEANS of the artist’s still can be. Because they already are. This is not just their intellectual property and creative copyright under threat as a group, but their future job prospects and opportunities to generate income and continue their projects in a supportive economic environment. Evidently, that is at risk? Why didn’t the makers of these products establish consent before running the work through a program that they knew would be able to pinpoint and recreate their styles and idiosyncrasies, even if only partially (though in some instances, the resemblance is unmistakable).
Let’s go a step further and say that there is an art to almost every field that requires in-depth study. Let us focus here because it is more expedient than debating the classist undertones of an analysis that might suggest there is a better area for the exploration of AI’s potentials to experiment and make waves and it might be jobs that are either degrading, dangerous, or for whatever reason do not pay well with no likelihood of paying well. Either way, people lose jobs, so let’s set that aside. Is there not an art to programming that AI could displace, making coders redundant? Is there not an art to writing an academic paper that is also one of thinking critically, minding the nuance in specific word-choices, syntax, source-material selection, collation of bibliography? If every individual’s sense of purpose is derived from Youtubing, this would be one thing. But that’s not the case.
It does matter what we choose to do in this moment, and what we demand of our legislators around the deployment of AI. So that thousands of employees are not suddenly laid off, with companies cutting workforce by 30-50%, with next to no warning, because of A.I.-adoption (as has been the trend for the last several months in the tech world).
I see this as a critical precipice and a moment of choice bearing some consequences for the future. How are we going to make sure that this technology does not displace, not the role of the artist -which is unlikely to change - but the means of acquisition that artists have option to pursue? Jobs with companies as graphic designers, story-board artists? As muralists and illustrators for tattoos, comics, music videos? Nobody knows just how quickly this technology can develop but what happened with Dall-E and is happening now with Chat GPT indicates it can move very quickly. Who is to say that in 5 or 10 years it will not have the capacity to be used by one guy at a company to do the work of dozens of labor-hours of the artist? Will foundations now make grants for prompt-writers, I wonder. And if so, for what computer projection on a wall that we haven’t seen a hundred times already with a little bit of side-eye….however, even artists making expedient video art have the good sense (many with complex, sophisticated cultural relationships to their subject matter and personal knowledge of the history of video art), to laugh about their unusual predicament and position. Algorithms cannot. Digital artists and their employers can acknowledge the hard-earned success and growth of artists that work in more traditional mediums. These algorithms cannot.
I say we not doubt the intensity of humanity’s worst quality - a greed that at the top of organizations seeks to maximize profit usually at the cost of labor first. AI can reduce the labor hours for a project, but will it be the artist or the project managers who wields control of it’s application in a heirarchy or team environment? At the same time, I hope we do not doubt our ability to choose here to do the right thing and codifying our ethics on this sooner rather than later.
Other questions this brings to mind, when one considers the vast number pf photographs of works across every medium that exist on the web and could be used by future programs if the lawsuits now being brought are not won by those seeking damages and clarifications. Would a bad outcome eventually lead artists working in traditional mediums to abandon their favorite materials in favor of computer programs that can only recreate the actual sensations of natural materials that so much of art involves? Maybe I am biased because I have my preferences and I like my ink and paper and strings. But if you’ve got an idea for something that now can be carried out by a machine with the right language-inputs, you are talking about possibly hundreds or even thousands of kinds of artists ( how do we put a number on kinds of artists? Is not every artist their own kind of unique?) that can become if not irrelevant, then exploited, if no restrictions and corrections are made to the trend as we have so far seen it play out. Many many artists who have already had their work fed into the ai machinery without their consent that did not get a choice as to 1.) how their work would be used, (2) who would be enriched by it, and (3) for the sake of what projects it was copied and fed into. Nevermind the effects on other people down the road in the future, of this manner and way in which art is offered up for consumption…mauled and torn-apart.
This brings to mind a lot of other more complicated questions: can an AI algorithm assist in the commitment of fraud? I think the answer, whether intended by its programmers or not, and as exemplified by the two lawsuits now in place against some of these companies that I know of (see here, and here), is yes. Regardless of how the jury or judge decides the cases and on what criteria they base that judgement, the fact of these cases being brought at all is an indictment of the projects concerned. It is behavior characteristic of so much of Silicon Valley and other ai-heavy, high-income, knowledge-based economic institutions-without-borders - act first, think later. Rather than ask for permission, ask for forgiveness. Only in this case, how could they even make time for all that they might apologize to?