Over the past few days I’ve found myself mulling the question of whether AI will destroy art and literature. Initially, I found myself comforted by a thought, articulated by Carrie Jenkins on bluesky, that since the value of art lies not simply in the product but in the process of its creation, art will survive intact. When I contemplate Van Gogh’s Starry Night, I’m not just considering a decontextualized pretty object such as an AI might produce, but something that results from human intention, contemplation and struggle and which flows from a life and its roots. So far, so good.
I was moved to think of Marx’s contrast in The Results of the Immediate Process of Production between Milton, who “produced Paradise Lost as a silkworm produces silk, as the activation of his own nature” and “the literary proletarian of Leipzig who produces books, such as compendia on political economy”. (See Capital vol 1, Penguin edition, p. 1044). The literary proletarian may be threatened by AIs, which can churn out such compendia, or perhaps boilerplate romantic fiction, but a Milton is not. But on further reflection, I think this is a mistake. Not that “Miltons” will entirely disappear but they will be oddities, isolates, like Sabato Rodia who built the Watts Towers.
The thing is, people do value products for their instrisic characteristics, divorced from the histories of the creation and creators. When people go to IKEA to buy a nice lamp or a rug, they are mostly indifferent to who has produced it: they want something that looks good, is affordable, and works. And AI can produce this, thereby depriving thousands of equivalents of the “literary proletarian” of their livelihoods. Sure, a few people might pay a premium for an Anni Albers-designed rug (and more for an original), but most will settle to adorn their home with an AI-produced knock-off at a fraction of the price.
The trouble is, that the elimination of the literary proletarians doesn’t simply leave the Miltons standing, unscathed. Mostly, art does not just emerge from a random genius popping up and producing great works but from a milieu which provides a context and an infrastructure. A network of other producers but also critics, dealers, suppliers, teachers. I believe Howard Becker writes about this in his book Art Works, but though it is on my to-read shelf, I have not yet done so. Some of those people produce output that is “art”, but since “art” is a prestige category, many others produce work that fails to rise to that level but which is merely decorative or entertaining. Many of them will have been trained in art schools or universities but have failed to make it as artists, but without them those schools will become unviable. In short, withouth the wider group of near-misses and engaged supporters it is hard to see where many artists will come from: thanks to AI they will lack a sea in which to hatch and then swim.
{ 30 comments }
Phil 01.02.26 at 9:51 am
I used to know a guy who was a former cabinet-maker. It’s probably true at some level that there will always be a market for well-made, artisan-built bookcases and chests of drawers, but it’s also true that a lot of people will settle for the poorly-made alternative if it’s cheap – and if you can pay a bit more and get something reasonably well-made (the niche that IKEA carved out), the market for hand-made furniture is going to get very small indeed – too small to sustain the trade of cabinet-making.
Perhaps it wasn’t Skynet gaining sentience we needed to worry about!
Lisa H 01.02.26 at 10:05 am
There will likely continue to be people who enjoy the process of creating art and get good at it over time, and who do it as a hobby, don’t you think? In that sense, many art forms might go the way of knitting, lace-making, and other traditional forms of arts and craft. Whether that can produce Miltons, I don’t know. It will probably depend a lot on how much free time people have… if AI could be used to reduce working hours, allowing people more time for hobbies, then who knows. Unfortunately, I don’t see this coming, at the moment…
Ethan 01.02.26 at 10:35 am
Not to sound all relativist here but the most important part of the question for me is considering how the valuation of art is contingent. Furthermore, how we value art I think is more about what we do with it rather than what we think of it.
As you say with the IKEA example, there is a certain pragmatic element to the valuation of art. If it does the job, it does the job. You can even say this of literature, film, visual arts etc. Tolstoy wrote that the definition of art is that which imitates and evokes a certain emotion in the audience. There’s nothing about the context of its creation which is necessary for doing so, which is why AI may suffice in many cases.
I’m also reminded of Bridget Riley arriving in New York, horrified to see her paintings reproduced on clothing and deprived of its “purity”. The base is the augmentation of constant capital, art mechanised and reproduced whether through AI or furniture factories, and the superstructurual effect is indifference to the context of the art’s creation, so much so we often don’t consider clothing or furniture as art, and other forms of art will probably follow that pattern.
Duncan 01.02.26 at 11:21 am
Human intention arguably needs to be present for art to exist, but that could take the form of interpretation. The reader, in effect, creates the art object. (Think of modernist experiments with algorithmic art or found objects.) Unfortunately, AI is eroding our ability to read and focus, too.
mw 01.02.26 at 12:38 pm
I don’t pretend to know how all of this is going to shake out. Will people demand that there be real artists behind the works–flesh and blood humans who can go on book tours and do publicity interviews and whose glamorous exploits and absurd foibles can be followed in the tabloids? Or will they be satisfied with so many Sophie Winters as pop stars?
One thing that I will note is that a lot of fields of art weren’t doing well long before the advent of AI. Compared to, say, 1950, a far smaller percentage of educated adults can identify a famous poet or painter, and literary novels and even theatrical films seem to have likewise been in a long secular decline. Sales and interest in all of these categories have been in decline for decades.
But there is some reason to hope that AI could offer new avenues of renewal. For example, it is conceivable that a single artist could soon produce a full length feature film, shot-by-shot, scene-by-scene using from an original screen play using Veo or Sora the way an author now writes a novel. The ‘actors’ and ‘locations’ would be purely imaginary in the AI produced film, but ultimately why should that bother us more than imaginary characters and places in a novel (or more than green screens and CGI–or matte painting, for that matter)?
Perhaps, too, people (ordinary, non professionals) who have visions for works of art but formerly lacking the skills, time, or money will be able to realize those visions with AI? Think of students studying art history with growing knowledge of historical styles giving them a palette they can then use in AI prompting. Maybe in the future, people won’t decorate their homes with prints of known famous art works but rather with images they’ve generated themselves? People once had to hire portrait painters (which only the wealthy could afford) and then professional photographers (within reach of ordinary people but very expensive) before they could capture rare images of loved ones. Then of course, DIY photography became possible (but still costly and requiring expertise) and now we are at a place where it has become ‘too cheap to meter’ (and not just for the industrialized world, but globally). With AI, aren’t the visual arts and music now set to follow a similar trajectory?
Personally, I’ll be much more interested to see what happens with the arts in the era of AI than I would have been to see recent (arguably tired) trends in art continued indefinitely into the future.
novakant 01.02.26 at 12:41 pm
people do value products for their instrisic characteristics, divorced from the histories of the creation and creators.
I don’t think this even applies to interior design, let alone art. Even the IKEA branding highlights the design process and their Scandinavian identity. And you find this appeal to originality in all sorts of other products like food, cars, clothing etc. Of course there’s always a lower end where bare necessity and price rule, but most people are able to tell the difference and aspire to something that is higher-end and more unique (this type of branding drives the whole fashion industry) – that’s why marketing departments are constantly highlighting these aspects of the products.
When it comes to the visual arts, the story of the artist and the message are everything and the end product is worthless without it. In film and TV, even the average Netflix user is very sensitive to the commodification of storytelling and they turn off quickly. They want something original, edgy, preferably driven by stars. There’s a whole part of the media industry providing making-ofs, biographical sketches and gossip even at the lower end of the market, without which much of this stuff wouldn’t sell. People want to be able to identify with the creatives. And few want to read an AI generated novel or memoir, they want to know about the author, go to book signings etc.
So I’m pretty confident that the human need for a personal connection and differentiating oneself through knowledge and good taste means that AI is less a threat than many think.
Laban 01.02.26 at 12:43 pm
@Lisa H 3 – I see Chris’s point though – if AI novels swamp literature, might it not discourage people from writing, and picking up the what in another context James Dyson calls “process knowledge” – the learning-by-doing which hones skills?
OTOH I hadn’t written anything longer than a couple of pages since uni days until the internet and blogs…
Laban 01.02.26 at 12:58 pm
In a slightly broader context, it might be instructive to find a long-established school without too much change in its intake – say a Kent or Gloucestershire grammar – and see if there’s been much change in the standard of writing in the school magazine over the years, especially in the smartphone era.
mw 01.02.26 at 2:55 pm
novant @6
“People want to be able to identify with the creatives. And few want to read an AI generated novel or memoir, they want to know about the author, go to book signings etc.”
That may prove true, but if so, we are all but certain to see humans presenting AI work as their own. At one end, this may be in the form of established writers or artists asking AI to produce new works or ideas in their own style which they then tweak and finish. At the other extreme, we’ll likely see new AI versions of Milli Vanilli or Mr Brainwash. In fact, I think we have to assume that right now artists of various kinds are already surreptitiously leaning on AI.
engels 01.02.26 at 8:14 pm
I’ll be much more interested to see what happens with the arts in the era of AI than I would have been to see recent (arguably tired) trends in art continued indefinitely into the future.
It kinda sucks then that all AI does is continue whatever trends are in its training data indefinitely into the future.
somebody who noticed who actually is forcing ai art onto the world 01.02.26 at 10:17 pm
It’s only the wealthy and powerful computer boys who want AI art to replace human art. Their sickened, rotten, repulsive minds could no more create art than they could flap their arms and fly to the moon. So when they see a 14 year old kid drawing something in her Trapper Keeper they want to make a computer to steal everything she’s done and do what she would never do for them – make something cool.
MisterMr 01.03.26 at 1:09 am
As a hobby, I draw webcomics. In the webcomic community everyone hates AI as the devil, for the reasons outlined in the OP.
From the point of view of art as self expression, I’m the same thing of Milton, and like me are thousands of webcomic authors, webnovel writers, hobbyst musicians and painters etc., who never were going tò make a dollar from their creativity anyway, even without AI.
From the point of view of capital A Art, there are probably very few Miltons in the group, in part because of skill, which anyway is a very ambiguous and relative concept in art, and might well be surpassed by the AI.
But largely the reason that I’m not Milton Is not skill, but cultural relevance, that largely depends on society and if society sees the creativity of a single artist as representing society in general, and not just that single artist.
I think the OP suffers from mixing the concept of art as self expression (that is independent from success and skill) with the concept of high art, that is a very small subset of that creativity.
I can imagine, for example, a situation where people just get tired of AI slop because there is too much and turn to hobbyst products like my webcomic even if they have a lower level of skill than AI, and yet since there are too many of those body can turn those into a job; or maybe the enormous amount of slop that can be producer in a short time will give more importance to gatekeepers like reviewers, critics or influencers etc..
Anyway, I don’t think the old model was that great for creativity, since a really large number of people has creative urges and only a very small percentage can make a living of it.
@mw
What is the “vision” for a work of art exactly?
For example, in a comic, do you consider the story part of “vision” and drawings part of “skill”? And what about dialogues, choice of point of view and lighting, etc.?
My opinion is that the limit is very fuzzy, and the idea of someone who has the “vision” but not the “skill” is misleading, the vision and the skill are largely the same thing.
Kris McCracken 01.03.26 at 4:35 am
A really sharp piece. The point about the “literary proletariat” as infrastructure rather than dead weight is the bit that stuck with me.
One possible extension: the real danger isn’t just cheapened outputs, but the collapse of apprenticeship. Once the paid, imperfect, learning-by-doing middle disappears, skill formation becomes sporadic and amateurised. You can still have obsessives and outliers, but the ladder itself is gone.
Culture doesn’t die from lack of admiration. It dies when the processes that make competence possible are no longer economically viable.
Alex SL 01.03.26 at 8:27 am
That all sounds like a plausible scenario. The same may apply much more immediately to software developers. This is not my area, but from what I am reading, and to the degree that I do know software that seems plausible, generative AI cannot (yet?) replace software developers. But it can (supposedly) replace the entry-level job for a future software developer, a boilerplate coder working under the mentorship of a software developer. If genAI is successfully implanted everywhere, the future looks dire not only for artists across music, writing, and visual arts, but also for software developers, translators, copy writers, and likely even any qualified profession. Again, not because any of them can actually be replaced by genAI, but for the reason outlined in this post: the entry level “they are still learning, but it is good enough for now” is going to be replaced, so nobody learns how to do it properly, so a generation later everything is broken.
This is, then, one of the potential trajectories we are looking at. Generative AI takes over => nobody learns how to do things properly anymore => civilisational collapse, both in the purely technical sense of becoming unable to produce software functional enough to keep our IT infrastructure running long-term and in the spiritual sense of largely ceasing to create any art, writing, or other expression that deserves to be called ‘culture’.
However, there are two other potential trajectories. First, this all largely goes away because it is too expensive to run. The prices people currently pay to use genAI are heavily subsidised in the hope that (a) inference will magically get much cheaper, which doesn’t seem to be happening to the required degree, and/or (b) even if inference stays expensive, everybody gets so used to having genAI that they will be willing to pay thousands of dollars per year in subscription fees, and/or (c) even if too many people hate genAI, the models are becoming so embedded in software everybody has to use (Windows, Office, Android, etc.) that they are forced to pay thousands per year to be able to exist. We will see how that works out.
Second, this is all irrelevant anyway because, whatever the price of inference with genAI, we are clearly not doing anything about CO2 levels, so in 200 years we have true artists again because the next dark age has about the same tech level as the previous one, and genAI has gone the same way Roman aqueduct maintenance went in 800 A.D.
Moz of Yarramulla 01.03.26 at 9:36 am
We’ve already seen this with literature, the explosion of fan-fic and online fiction followed by Amazon largely taking over the small publisher market and then Amazon vomiting AI slop all over the fiction market (arguments about the people running Amazon vs the people using AI ‘investment’ to generate fiction are correct). Finding an actual book written by the actual author increasingly means following the author on social media and tracking their releases. Or hoping the physical bookshops are doing that task for you.
Similar things have happened with painting becoming photogrpahy becoming photoshop becoming AI image generators. You can get anything printed and framed, right up to 3D printed “real oil texture”.
A lot of people don’t care where the art comes from, or whether it’s really art in any meaningful sense. The thing, being it physical or digital, is a tchotchke with little significance, at best a memento of some minor event, at worst a decoration. Like lines in a Linkedin resume they serve merely as prompts encouraging (or discouraging) interaction.
What’s notable to me are academics like Quiggin and Gruen going all in on AI. Writing articles, illustrating them, to the point where I’m not sure whether Guren’s human input was limited to “write another missive in my usual Substack style”.
mw 01.03.26 at 10:40 am
engels @10 It kinda sucks then that all AI does is continue whatever trends are in its training data indefinitely into the future.
AI diffusion models seem capable of creating novel combinations of ‘influences’ from their training, as many human artists do. But perhaps you’re right that true novelty in arts will be something that only humans can produce, in which case fine — that’ll remain a niche for humans.
MisterMr @12 For example, in a comic, do you consider the story part of “vision” and drawings part of “skill”?
Not necessarily. As an thought experiment, imagine an experienced cartoon artist who becomes a quadriplegic. They are no longer able to draw, but they could train the AI on their existing work and then use their voice to describe what they ‘see’ in their head. Do you think such an artist could continue to produce good work this way? If not, why not?
engels 01.03.26 at 1:47 pm
I agree with MW that the problems predate AI. Corporations are in many respects functionally similar to AIs and the culture industry has been flooding the world with generic slop for a century. Then there’s the credentialling and routinisation of traditionally creative fields the PMC has effected over recent decades, clearing the ground for what’s to come: Uber, but for BBC New Generation Artists.
Dave 01.03.26 at 3:41 pm
I wouldn’t simply grant that AI can produce even one instance IKEA-quality literature, let alone produce an IKEA-quality literary ecosystem. I don’t hold pop fiction or other popular artifacts in especially high regard, but AI isn’t even up to that. AI turns out garbage at scale. That’s not what pop lit is.
engels 01.03.26 at 6:28 pm
A lot of people don’t care where the art comes from, or whether it’s really art in any meaningful sense
True, but they probably never bought paintings or reproductions but cat pics and Live Laugh Love signs.
MisterMr 01.03.26 at 6:50 pm
@mw 16
I can speak just from my experience, but I already claimed to be Milton, so I’ll go.
Premise: I’m just speaking of “creativity”, not about “quality”.
If I try scripting a story, or drawing a picture, I’m not starting with everything all already clear in my mind.
Rather, I’ll start with some ideas of Who the characters are and what they want, and a few main plot points, or with a simple sketch of maybe the heads of the characters and their expression.
Then as I put this stuff on (electronic) paper, looking at It from the outside prompt my mind with more details, and so on and on, until the final result is way more detailed than the initial idea, and often quite different.
This “processo” is what is generally called inspiration, or why many writers claim that characters have a life on their own.
It basically means that a lot of the production process Is unconscious.
If you cut out too much of this process, the result will be maybe a fun idea represented very well from the AI, but without this dept that is die tò the process, IMHO.
This again is just about creativity in general, the point of view of the author; then if you ask why a certain product is commercially succesful, or why is it called high art, these are very different stories.
Ray Vinmad 01.03.26 at 8:42 pm
Visual art that is manufactured or automated already had a huge market. There will always be manufactured and automated art. Machines could always make paintings over and over. The humans that designed such paintings for decor are not of interest. The ability to remove agency from the process of something aesthetically pleasing has been in place for some time now. AI is simply something which is similar to this but which can be tailored to someone’s preferences perhaps more individually —like ‘make the Mona Lisa but in only red and white.’
So how could AI could change this for the world of art-making and art criticism? True, the process of manufacture has come into the art world in various ways with certain riffs on it. But it is only effective when Jeff Koons or whomever does it when it is attached to his intentions.
Visual art that people want to think about or talk about requires a human creator, a being with intentions, life experience, a history, that can see and feel and suffer as the viewer sees and has emotions not merely repeating words signifying emotions. Otherwise, there is little or nothing to talk about.
We can enjoy the aesthetics of a rock that is visually magnificent but we would not discuss the intentionality involved in the rock’s visual magnificence. It is made by happenstance. Any sort of art-like creation—rust on a sign does not elicit the same commentary or reflection that occurs when an artist uses rust.
One could talk about how the rock came into existence. That is very interesting as a subject—but it is not what we do when we experience art.
The Eliza effect might confuse people to some degree about this —or they could be intentionally confused by giving a bit of Gen AI a name and reporting its words in an emotionally-laden way. Then you would have ‘an artist.’
But there would be pretense or delusion in this. People might enjoy it—but it cannot replace what art is now. It would not BE what art is now. It would be analogous to automation or manufacture.
Also, humans continually make things which are aesthetically pleasing to them. It is possible that the MARKET for art would drop out. It is possible that there would be no intellectual discussion about art because of some calamity. But it is extremely unlikely there would be no aesthetic making because every human culture does this and sometimes in great quantity whether or not it is appreciated or lauded or commercially viable. We can be distracted from doing it in our daily lives by manufacture. Probably fewer people in modern society sing and dance and make music and make beautiful textiles and pots and so forth —at least not as often. But to get people to stop doing this you generally have to make laws against it and punish us for doing it. It’s just something humans as a species always do—not every individual human—but in every group people are there who will do this.
So I don’t think art can die out.
SusanC 01.03.26 at 9:02 pm
much of the AI generated content I encounter on the net is really poor ( “slop”)
Two interpretations
A) well, maybe the AI will get better
B) we really notice when the creative element is missing — it is actually hard to make good art using AI
If it turns out to be (B), there is some hope for artists continuing to exist. Though their skill may lie not in painting or drawing, but in knowing what to create
SusabC 01.03.26 at 9:09 pm
We might compare photography, where any fool can press the button on the camera and it creates a technically correct copy of the light falling on the lens, but
turns out, composing a photograph still involves a considerable element of skill
Hidari 01.03.26 at 9:46 pm
Unfortunately I read a lot of threads on ‘X’ which I then don’t bookmark and forget about, but one thread I remember reading relatively recently was about how American TV had its Golden Age in the 1st 2 decades of the 21st century precisely because it used writers who had grown up writing in the 1990s, the 1980s, even the 1970s and had really been taught how to write for TV. In other words, they had been taken on as ‘young cubs’ and then given larger and larger things to write, but still on a production line, for stuff that basically…wasn’t very good (detective series etc.). For example David Chase worked his way up till he was writing regularly for the Rockford Files and other things of that ilk….such that when he was given the freedom by HBO he knew all the ‘nuts and bolts’ of how to write realistic dialogue, how to ‘keep the viewer watching’ and so on. It was precisely because The Sopranos wasn’t his first ‘rodeo’ that it was so good. And now that is kinda fading away and it seems that the Golden Age is drawing to a close.
Now, this ties in with another thread about AI and its effect on the workplace. What this person (forgot who) was saying was that AI is not leading to mass redundancies, and people at the very top are not losing their jobs. What are being replaced are the smaller, less ‘sexy’ jobs at the ‘bottom’ which can now be easily automated.
The problem is that to get to the top you have to work at the bottom first, and work out how to do your job from first principles, on a sort of conveyor belt system. But these jobs are increasingly scarce, and we are starting to see the effect of that now, with fewer and fewer artists, writers, musicians ‘breaking through’ because the ‘dog work’ jobs (writing for adverts, writing muzak, writing bad sitcoms for late night low audience satellite TV companies) no longer really exist any more. Why? AI.
So artists are not getting a chance to learn their craft: nor are they getting a chance to get a first foot in the door which may lead to greater things. And as AI improves, this situation is likely to get worse. This fits in with other things that are intrinsically good (like increased longevity for example) but which might not be good for art…like (e.g.) The Rolling Stones still releasing albums and going on tour etc…..plus the internet making available almost the entirety of the arts for anyone who wants it…why bother creating anything new? Especially as, as is well known, there are high, and increasing, pressures on the ‘cost of living’ at the same time as it is getting more and more difficult to make a living as a musician or an artist (because of the internet and the way it erodes copyright)…commercial pressures that AI is probably going to increase (why bother paying for background music for your sitcom or play when you can just ask ChatGPT to do it?).
So it all seems kinda bleak and I think that we are starting to see the effects of this now. The 1st two decades of the 21st century were hardly an artistic Golden Age but enough good stuff was produced such that you could say that the ‘tradition’ was being kept alive….but…the 2020s?
Maybe not so much.
MisterMr 01.04.26 at 10:28 am
@Susan C 22
There is also C) if I can churn out 100 “aristic” images/texts per hour why should I worry about quality, I’ll just cast the net wide.
mw 01.04.26 at 2:40 pm
SusanC @22
Two interpretations
A) well, maybe the AI will get better
B) we really notice when the creative element is missing — it is actually hard to make good art using AI
There’s a possibility C as well — which is that you don’t identify better AI generated art as the product of AI at all. For example, what is the giveaway that this video clip is AI generated (other than it would be really difficult to catch this moment in the real world with a camera?)
Katherine 01.05.26 at 4:08 pm
@Hidari This is true also of any number of other professions. Take my own ex-profession – law. It is undoubtedly the case that AI could do a lot of the grunt work that used to be done by trainees – drafting, research, that sort of thing. Still needs checking, but the churn can be done at much faster speed. Trouble is that you need, as a trainee, to work your way through this grunt work in order to develop the drafting and negotiating skills that you need to do the stuff that AI absolutely can’t do.
Stealing this from someone else; all these professions used to be triangles and now they are diamonds. Before long there won’t be enough people who were at the bottom of the triangle to take over as the middle, let alone the top.
engels 01.07.26 at 1:20 pm
you need, as a trainee, to work your way through this grunt work in order to develop the drafting and negotiating skills that you need to do the stuff that AI absolutely can’t do
Not to pick on this but it’s related to the argument of the post and I’m wondering… do we know that the amount grunt work you used to have to do was optimal for partner formation? Wouldn’t that be a rather lucky coincidence?
Tangentially, why have junior lawyers’ salaries skyrocketed in the last few years if the number of those roles is shrinking?
engels 01.07.26 at 1:27 pm
It seems like the argument that with the advent of calculators people won’t learn arithmetic, and you need to learn arithmetic to be a serious mathematician… to which the solution is: make aspiring mathematicians do arithmetic (just not as much as before).
MisterMr 01.07.26 at 2:47 pm
@mw 26
But the problem is not if the content is of technically high quality, but if it is meaningful.
So we’ll have a lot of content that is of high technical quality, so that if you look at it at first you’ll have a “wow” effect, but that is done with small effort and often with small inventivity (and se we’ll have really a lot).
So in this sense weìll have a lot of high quality slop, that at first glance can’t be distinguished by meaningfull stuff, and therefore will drown the meaningful stuff.
Something like this happened to the DeviantArt website (a very famous community for illustrators). Fundamentally it now has been deserted by most really existing artists because AI production swamped everything, see this reddit thread on it for example:
https://www.reddit.com/r/DeviantArt/comments/11nz4rf/has_anyone_else_noticed_the_rise_of_ai_art/
Comments on this entry are closed.