Friday 11 August 2023

Still Satanic After All These Posts

Like a 1950s coal miner in a County Durham pit village, I return once more to the tap: yes, I am writing once again about AI art. (See posts here, hereherehere and here.)

This time, the trigger is a comment left on one of these posts, which raises important implicit questions that I think need rendering explicit and answering (if I was a posh French post-structuralist I might at this juncture invoke the magic word 'problematise'). I hope the person who left the comment doesn't take this personally; differences of opinion are what make the world go round. Anyway, here it is in its entirety:

I think there is an overlooked point with a lot of AI art that is much the same as when people get outraged about piracy: it's not a lost sale if the person wasn't going to buy it anyway, and equally it is not a lost commission if the person wasn't going to commission art for the project anyway. But it does mean that person has a thing they otherwise would not have had. Someone using AI art to make a whole bunch of character portraits for their home RPG is inarguably better off with AI art than if they are doing without.  

This obviously overlooks plenty of other moral complications, but democratization of information and art and other things is usually a good thing. Usually.

This is an instructive comment because it rests on a number of assumptions which people who are in favour of AI art, or see nothing wrong with it, tend to make. The first of these is that it is 'victimless' in the sense that it is not literally taking the bread off the table of any starving artists - all it is doing is giving people who would not normally be in a position to buy original illustrations the opportunity to make 'their own'. The second of these is that art is basically akin to a consumption good - the more that one possesses, or is able to have at one's finger tips, the better. The third is that AI art represents 'democratisation' in the sense that it makes art more accessible. And the fourth assumption - let's call it the meta-assumption - is that AI art can be described as 'art' in the first place.

The first assumption is easily 'problematised': a strict utilitarian might agree that since no identifiable individual is per se harmed by an algorithm spitting out pictures of Optimus Prime in a Starfleet uniform or Yoda replacing George Costanza in scenes from Seinfeld or whatever - no artist's work is being replaced! nobody is losing money! - then there is nothing wrong with it. But (even if we grant the premise, which I don't -  individual artists will surely at the margins lose out on commissions, and in vast quantities) strict utilitarians are always, rightly, derided for failing to see the bigger picture. Even if it is true that the proliferation of AI art does not hurt any one person in particular in pecuniary terms, it is still perfectly possible for it to cause something which we love and cherish to be denuded of value and distinctiveness and for our lives to be diminished as a consequence. A good analogy here is music: we now have more access to more music than ever before, at the click of a mouse button, but I think it is genuinely an open question as to whether we collectively enjoy music more in 2023 than we did in 1993. I used to own vastly less music than I can now listen to through Spotify, but I can't honestly say that I listen to it with anything like as much enthusiasm as I once did - and that's not a factor merely of age, but of the fact that there is now so much of the stuff available and of the fact that it is so very accessible. I can barely listen to a track from start to finish without my attention wavering to whatever is coming up next on the playlist, let alone become lost for weeks inside an entire album, listening to it over and over again, as I would have twenty years ago. Without going too far off on this tangent, I think the point is made: we can all recognise that a phenomenon can be 'victimless' while still making our experience of the world worse.

The argument against the second assumption follows from the first. Our commenter makes the (to my eye, highly tendentious) claim that 'Someone using AI art to make a whole bunch of character portraits for their home RPG is inarguably better off with AI art than if they are doing without [emphasis added].' Well, 'inarguable' is a strong word; I think it is really very arguable indeed. Is someone 'better off' if they have more of a good per se? It depends what 'better off' means and what the good is. This person using AI art to make a bunch of character portraits may be better off in the sense of having physically more art, but is she better off than a person who worked hard at learning how to draw so that she could produce her own character portraits, or a person who asked his son's friend, a talented amateur artist, to do some pictures and thereby encouraged the kid to take his work more seriously? Or is she even better off than somebody who is forced to rely entirely on his or her imagination? These, too, are open questions. The argument that more is intrinsically better needs more justification than simply saying it is inarguable. (And this takes us back, of course, to the music analogy made above.)

The third assumption, meanwhile, relies on another problematisable assertion, which is that the widespread availability of AI programs which can create plausible visual images to order for trivial cost makes art more accessible and hence will 'democratise' it. Now, nobody could of course dispute that the phenomenon of AI art will allow people to make almost infinitely more pictures than ever before, and of infinitely more variety. But this raises the important question of what accessibility and democratisation actually mean. Is outsourcing creative work to an algorithm aptly described as 'democratisation'? It doesn't feel that way to me. If anything it feels rather like the opposite of democratisation. Democratisation of art conjures in my mind increasing opportunities for people to create their own artwork and to receive an education in the history, philosophy and techniques of art, but if the use of AI to make pictures can be described as anything, it is surely not any of those things. Quantity has a quality all of its own, as the saying goes, but is having the ability to command a robot servitor to make billions of pictures per second really going to give anybody who did not already have the means available the opportunity either to get good at making art themselves, or to learn anything about the history of art and what that entail? 

Hovering over all of this, of course, is what I earlier called the 'meta-assumption' - which is that AI art is in fact 'art' in the first place. Is it? Or ought 'art' to mean something that is created by human hand? And is pastiche, which is essentially how AI operates - through magpie-like assemblage of prior images which are then glommed together to make things that are 'new' - to be understood as 'art'? Reasonable people will differ in their responses to these questions, and I am not about to launch into a full-scale discussion of the meaning of art here, but the point is sufficiently made: to call the phenomenon of AI art 'art' itself begs the question, and forecloses sceptical inquiry. If the assertion is accepted, a whole load of assumptions follow, and if it is denied, an entirely different lot of assumptions is made. 

More broadly, it increasingly seems to me that what is at stake here is much more fundamental than the issue of what happens to actual human illustrators and whether they will lose careers and livelihoods. Questions such as the meaning of art, the relationship between art and artist,  the distinctiveness and importance of humanity, the nature of man as distinct from machine, the artist's moral right over his own work and so on are now all, clearly, up for grabs - and reveal themselves to be genuinely divisive matters. Put bluntly, whatever one thinks are the answers to these kinds of questions, one's views seem to derive from coherent ideological frameworks that do not significantly overlap. This does not bode well for future cooperation - and suggests further fragmentation of the hobby as those for and against AI art go their separate ways. 

46 comments:

  1. I've always been a highly visual person. Often with a mind for creating things that my talent and skills don't support. I have enough hobbies and skills, I'm quite comfortable producing visual "art" by slapping together things others have made - directly or indirectly moderated via "AI". I don't care if it's "art", it provides me with enjoyment, and sometimes useful results.

    Since heraldic devices are my most specific joy, "AI" can only take me so far (for now - I'm working on that). But the marketplace for such work is small and in my experience snotty. Even when I've tried to throw money at people to help me realize my vision, they've refused because I'm not doing "real" heraldry. Although credit to my favorite creator in that space, who refused not because he didn't want to take my money (he seemed enthusiastic about my idea), but didn't wish grant me the license I asked for.

    I'm not patient enough to engage with you at the level of your writing here, and maybe I'm not smart enough either - although I believe I am. I'll just say that I see a lot of conservatism and elitism here. And some of it is very difficult to credit beyond mere subjective opinion, you relationship to music for example.

    I don't find - so far - compelling any argument that the arts should be excluded from the dangers of new technology. My field is constantly at risk, and I won't deny that it creates fear in me; but history tells me that such disruptions usually carry a great benefit over time. The one concern is of course intellectual property rights - a case I don't find compelling because I don't have much sympathy for the rent-seeking we see in the creative spaces (or the corporate ones for that matter).

    Any disagreement aside, I'm always excited to read your thoughts. Thanks for posting them.

    ReplyDelete
    Replies
    1. I don't think I'm being elitist - in fact rather the opposite. The point I was trying to make was that I fully support democratisation of art if that means more people are able to create it, view it, and learn about it. My argument is that AI art won't facilitate any of those things.

      I get the argument about disruptions usually being beneficial over time. This point is often made. The only thing I would say is that just because previous tech disruptions have been beneficial in the long term doesn't mean this one will be. We're not talking about a law of physics!

      Delete
    2. If we look to the past, disruption in the form of spontaneous mass-production has precedent, but it is quite arguable whether they benefit the public or the arts. Take, for instance, the art of textiles. The invention of modern textile industry has certainly made textiles more available and, given their objective need, this is certainly good. But it is undeniable that vast swathes of knowledge of the art has been lost, and few people are able to make textiles as a living in the modern day. The textiles we do wear are mass-produced with cheap methods, nothing like the durable, custom made clothing of yesteryear.

      AI-generated images can be held to be quite similar. If that old game of Cheat the Prophet does not play its tricks on me, I foresee a generation without visual artists. Much has already been lost in the modern arts: we all know a digital painter, but few know a sculptor in marble. We shall throw away images as we now throw away clothes, where before they were held valuable. Man-made art shall become, even more than ever, the commodity of the elite, if the elite care for it at all (looking at you, elon). Not to mention that modern people already have more art than they can ever need. Indeed, if an AI can be trained to reproduce a style, this is in itself proof that such a style is quite populous without AI.

      I knew a woman from an african country (I forget which), which was poor in imported clothing, but rich in tailors. She is not a rich woman, but her attire is quite original, dazzling, and delightful, and is in great portion hand-made. But, she has said to me, her country picked up a fad for Rayon, Polyester, and other synthetics in some decade or other. Though her clothing is cheaper than it otherwise would have been, and remains quite durable (though not as durable as it might be), it is often uncomfortable. If I recall correctly, the price of good cotton has also increased, leaving many of her countrymen in fabric quite unsuited to tropical weather. It has, in some ways, been beneficial. But it is not *merely* or even *certainly* beneficial.

      Delete
  2. I don't recall whether or not I've made this point previously, but my take has always been that art is an 'action undertaken'. If the word 'art' is not precisely a verb, then at the very least it is a noun ripe with potential, loaded with the immanence of movement.

    The dichotomy between the experience of the creation of art and the experience of spectating *upon* art, I would argue, is the central wound in the soul of the modem 'west'.

    AI creations have merely brought that tension into the light, where I think it has a chance to transform/release. What is art if not the beautification of relationality? The true democratisation of art is in this realisation.

    A lot of baby goes out with the bathwater. But the potential revelation, is I believe, immense.

    ReplyDelete
    Replies
    1. But what is the revelation? Tell us!

      Delete
    2. Something arising from the following assertion:

      Art is only meaningful/potent in the act of its creation. And that any and all benefits to the self and the community, arise from this. Art as Spectacle is ultimately disempowering/severing antimatter, arising and densifying over progressive centuries of culture.

      This, I believe, is what our distinctly Neo-Platonic AI friends make explicit by the image of their labours.

      Delete
    3. Interesting perspective. I don't agree, though. I went to a local art gallery a few weeks ago to see the work of Spanish renaissance masters and left filled with the desire to create. Art as spectacle has a positive relationship to art as the act of creation! No?

      Delete
    4. Absolutely. Hence, the statement regarding throwing out the 'baby with the bathwater'. Obviously, I personally, have been inspired countless times, by countless works. Where would I be without Sergeant Pepper?

      However, consider the following:

      1. Art might be said to be the difference between doing something in a way that enchants and being blindly unconscious of one's actions. There is a way to drink a shot of whiskey that is artful and a way that is not. A toast, for example. The same thing might be said of any action, any activity. From accounting, to street cleaning.

      2. In this sense, there is nothing that is not Spectacle, since our experience is mediated by images. But this is arguably a side effect of art's existence.

      3. Within communities, wherein the inspiration for the creation of art is in the motion of the day, as opposed to the image of another's day, is where we can perhaps identify a crux. I sometimes identify folk art as being connected to time and place, which is another primary influence upon the art of one's life.

      4. The wide variety of indigenous cultures approach to art (many wouldn't even recognise such a separate entity), is also worthy of note. Art here, is the sacred in day to day life. It is doing things right, with care and with exactitude. I think of the Massana dance /play/drama/standup routine of the Mbendjele women :

      "Men who have been aggressive, disrespectful or even lazy lovers are playfully, humorously, but assertively, held to account by the women in the communal space...rather than entering into a combative space, they embody their own quality of collective power, expressed through potent laughter, song and sexuality. But the women choose not to hold centre stage for too long, as this is an egalitarian society, and not a hierarchical matriarchy. They believe that to hold the power for too long could lead to resentment, so they willingly allow space for the men to have their turn, through their ritual called Ejengi." Bruce Parry, Taiwai III

      This is art as human law. Not art that sits and elicits the desire to create in turn. Such an idea would be madness within this context.

      5. It is context then, which is challenged by the AI situation. Our framing of 'art' (no pun intended). The way we situate it, the quality of our relationship to it. What is 'it'? Because, by some definitions, it is life itself... or perhaps 'life as sacred'. Your original assertion regarding AI as Satanic, I think hold water, in the sense of Satan as the material universe in primacy, as the beginning and the end. "There is nothing but me". In a more prosaic way, it is simply watching something live life for you. Perhaps more galling is the notion that AI art, is essentially watching a mineral based language model living life for you.

      6. Ultimately,I don't know what any of it means. I don't know if I'm right, or merely wrong in an amusingly distracting fashion (or more charitably, 'less wrong') I return to my original words 'the potential revelation is immense'. There's an awful lot to still be said here, but hopefully this is useful grist for the mill.

      Delete
  3. Very thoughtful post. I definitely agree with you on the first point. Sure, people with no budget can use AI generators and get images to use for whatever, and technically this doesn't harm anyone's bottom line in and of itself. But those people aren't the only ones using AI generators. And if companies big or small turn to them instead of paying actual humans to create art, it does drive down prices and hurt the artists. As you say, the utilitarian analysis of a single transaction ignores the wider implications.

    I also agree on the second point. You point out the benefits of NOT relying on a computer algorithm to produce art. And even from the point of view of the commenter, that person doesn't need AI generators to produce pictures for their campaign. Google Image Search can already turn up all kinds of useful images, and may even be less stressful than trying to finagle the exact right description to feed the AI to get the image you really want.

    On the third point, I agree with your sentiment, but I view the idea of "democratizing" art a bit differently. Art is already democratic. Anyone can produce art, even if the best they can manage is a stick figure. Comics like XKCD have been chugging along for years now with nothing better than stick figures. This point is moot, IMO, because as I said, art is already one of the most democratic (equally available to all) things we have.

    On the final point, I'll leave it up to the art world to debate whether AI generated images are art or not. On the subject of pastiche, though, I actually had this thought on one of your earlier posts where you mentioned that all AI can produce is pastiche. Pastiche isn't all bad. Lots of young writers and artists get their start, and practice their talents, though pastiche. Of course, qualitatively, pastiche is lesser than original content, but it plays an important role.

    One of the best assignments I had as a creative writing major was to purposefully write a story in the style of a published author. I chose William Goldman, and tried to write a Princess Bride (the novel version) style short story. Of course it was no where near as witty, but I think I had a bit of the irreverent style in my story. It taught me a lot about my own style of writing. Pastiche has a place. AI generated pastiche, however, has no, or very little, value in teaching the next generation of writers and artists.

    ReplyDelete
    Replies
    1. Yes, I agree about pastiche as a learning tool for actual human beings. Actually, your comment reminded me of Hunter S Thompson, who used to actually just type out pages from his favourite novels (Fitzgerald, Hemingway, etc.) word for word in the hope that through sheer literal imitation he would thus learn to write great prose.

      Delete
    2. Hey, that's how I learned computer programming, by typing in programs from "Compute!" and similar magazines.

      Delete
  4. Aside from the fact that AI stands for "artificial intelligence," I know nothing about AI. I mean, about the state of AI these days or how to do "AI art" or use chatgpt or whatever. I don't give a shit about AI (even if I should) and have too much on my plate to bother myself with all this stuff. Probably the reason for my last pithy comment on your last post on the subject.

    I will say this:

    In doing various writing projects (i.e. self-published game books) the most difficult aspect is getting illustrations for the things. I solicit artists, I offer to pay artists, I DO pay artists, and...still..getting the artwork in a timely fashion for my publications is like pulling teeth. Even from the RELIABLE ("non-flaky") artists.

    If I could punch a few phrases into a computer and get a nice, clean, useful image spit out, my job as a game publisher would be O-So-Much easier. Because people expect art in these products. And I'm not a (good) artist. I can cobble together words, but the illos?

    Fuck. Relying on other people sucks sometimes, you know?

    Should AI illustrators replace real world artists struggling to earn a little money at their craft. No. Should it be easier to get said starving artists to do some fucking work in a timely fashion for an honest dollar? Yes.

    Is this a First World Problem? Absolutely.

    I still don't give a shit about AI (even if I should). I don't have the time these days to finish my latest book projects anyway, let alone chase down artists to illustrate them.

    But I can see how...if I DID, fortuitously, have the time and opportunity to finish one of my recent works...an efficient computer-generated art source could be a gosh-darn GODSEND to an indie publisher trying to get his/her work finished and on the market.

    Others may disagree.

    ReplyDelete
    Replies
    1. I feel your pain and you make what is I think the strongest argument for using AI art in nerdy elfbooks.

      Delete
    2. This is the key for me: "Because people expect art in these products." Why do people expect art? Perhaps because it is inspiring or useful at the table; we can all think of some pieces of game art that meet those criteria. However we can also see that the vast majority of illustrations commissioned do not meet this standard. Hobby publishers would be bucking the trend and courting bad reviews if they included only art that was useful or inspirational, because there will never be enough of such to meet current expectations. I argue this is because corporate publishers have set an illustration standard that only they can afford to meet in order to put competitive pressure on the smaller creators who cannot. The way they do this - paying writers by the word to generate content, and then not paying them anything extra to craft art orders for the illustration of that content which can be sent to whichever of a mass of freelancers will produce the described image most cheaply and reliably - seems to me designed to meet a bar for quantity of art, not its quality or utility.

      Delete
    3. Interesting argument, Tavis. Certainly I think the expectations about art derive from the 1990s splatbook era, when every RPG book you ever saw was chock-full of images.

      Delete
    4. Contrasted with the original post, JB's comment is pretty repulsive.

      We're all thinking high thoughts about the nature of art and well-being of artists, until we find ourselves on the other side of the transaction, and then suddenly there's a stark focus on how convenient it is to have a lever to get people to give you more of what you want for less.

      Delete
    5. To be fair to JB, it is incredibly frustrating dealing with creative people - who tend for whatever reason to be very flakey and unreliable and frequently find it impossible to get things done on time, if at all. I've learned to be phlegmatic about this (and also recognise that I can be just as bad, so it would be totally hypocritical of me to complain), but it is a big problem to anybody who wants to get a book done.

      Delete
  5. I imagine there will be a great deal of think piece-ing about this, but in three generations your stated position will have a constituency entirely made up of the sort of people who think medieval peasants had a better life.

    Organized, privelaged labor has a chance of extracting huge economic rents by perhaps impeding AI replacements (e.g., teachers, government paper pushers), but artists and writers are about to bifurcate into the very productive and the unemployed.

    As to being better off in some metaphysical sense? Who knows. It is amazing that we are so rich, so safe, so free. And yet we have antinatalism, euthanasia, sterilization of mentally ill youth as a civil rights cause - a veritable death cult. Why? I don't know. I am not at all facetious when I say Spotify, Streaming TV and this smartphone I'm typing on may be to blame.

    ReplyDelete
    Replies
    1. I'm dubious about your first paragraph because I agree so much with the third. I am increasingly of the view that we're on a downward trajectory that will fairly rapidly accelerate. I don't think things will get to medieval peasant territory, but we are heading for a serious decline in living standards. And the death cult that you mention combined with the technology is clearly a driver of that decline.

      Delete
    2. I don't know that we will have a decline in living standards by any economic measure. In fact I suspect the opposite. But ultimately the convenience of not having to deal with other people, and the ability to wander through life easily through soporific technology, will make sacrifice and care seem arbitrary.

      We have developed supernormal stimuli in ever more aspects of life, so we can lull ourselves to old age comfortably and distractedly, and avoid facing up to the sadness and wonder of being a human being for a brief period of time.

      Delete
    3. I think the long-term trends are almost all negative, but steering away from statistics about X or Y, my biggest concern is simply that our societies have become unmotivated to perpetuate themselves either culturally or biologically. A society in that position is heading for big fall.

      Delete
    4. On that point I very much agree. People scoff when a multimillionaire says "money doesn't buy you happiness", but in a sense many in our society are like trustafarians at expensive prep schools. Very little material hardship compared to 10,000k generations of ancestors, but we don't feel more special than our peers, and we don't know what to do without ourselves.

      We celebrate the bravery of women saying they regret becoming mothers because of the inconvenience of parenthood. It is sick, nihilistic and narcissistic, and spreads the emotional and moral poison that one should measure the joy of having a child against the excitement of 10 days in the south of France.

      Delete
    5. Yes, 'bravery' has come to mean saying very outrageously self-centred things. I don't know quite what that says about our societies, but it's nothing good.

      Delete
    6. Jeez. That got very dark, very fast.

      Delete
  6. I really enjoy the (usually) amateurish character portraits that people draw on their character sheets. That spontaneous expression of creativity is central to what I enjoy about this hobby. I feel kinda sorry for people who outsource part of the fun to a robot.

    ReplyDelete
    Replies
    1. My enjoyment comes from almost exactly the same source as the one you mention.

      Delete
  7. I really enjoy the (usually) amateurish character portraits that people draw on their character sheets. That spontaneous expression of creativity is central to what I enjoy about this hobby. I feel kinda sorry for people who outsource part of the fun to a robot.

    ReplyDelete
  8. I've all along disliked computer-generated anything in FRPGs. Give me hand-drawn maps in which every dungeon door and every tree in the forest is a little bit different than all the rest! None of this "point, click, and fill" rubbish. I first noticed this in, I think, 1981 (when I was 11 years old). Contrast the wonder of Tolkien's map of Wilderland, Darlene's map of Oerth, B2's outdoor map, and Kuntz's Bottle City dungeon map (all done by hand) with the soulless dreck of the D&D Expert rulebook's map of Karameikos or X1's continental map (both ruined by their ubiquitous little mass-produced symbols). I'd rather have an average 10-year old draw my D&D maps on scrap paper with a dull pencil and no eraser than have any computer have any role in making them.

    ReplyDelete
    Replies
    1. I almost agree, but there is something very charming to me about Hexographer/Worldographer maps.

      Delete
  9. Good rebuttal on Ai democratizing art. Its not actually making it easier to create, only making it easier to comission. Worse still, everyone is comissioning the same few entities instead of the world of human artists out there.

    If there was a way to let people create like rembrandt without the effort rembrandt had to go through it would be a good thing. (I dont think its the effort and suckage that makes good art good, think of the implications and what then we'd have to do for the best art). But AI is more like making a homunculus out of rembrandt and having it paint for you. Is commanding the homunculus creative work? A better question is why would you want to do that all day, whether its "artistic" or not?

    ReplyDelete
    Replies
    1. 'Is commanding the homunculus creative work?'

      That's a good way of putting it. People want to convince themselves that it is. But it's a simulacrum of creativity.

      Delete
  10. I don't think it will lead to further fragmentation - the battle lines are already distinct. In my lived experience, playing with a GM who has better visual aids is usually not more fun. Improvements in this area often come at the expense of flexibility, responsiveness to player input, and other qualities I value more. Who thinks more visuals are better? I find this is usually people who are selling visuals, or who have adopted the dictum that one's virtues come from what one purchases and thus need to align their views with the profit strategy of their favorite publisher. Looking at our own hobby is proof that the profit motive is rarely supportive of what we enjoy about being human. AI is just a new way to run the same algorithm people have been following since long before we were born.

    ReplyDelete
    Replies
    1. Good point - I'm very sympathetic to this kind of argument. The way I sometimes put it is that we get the technology we deserve. We will get soulless algorithms producing entertainment for us because we have been producing entertainment that looks like it was made by algorithm for some decades now.

      Delete
  11. Your last paragraph is the crux of it for me. With all of the things we could be using this technology for, why is it to take away the most human of endeavors: to create? Machines could do the drudge work for us and allow us more time to create art, music, etc. Instead we are heading for a future where machines make the art and we get to mop the floors? Why is that how it has to be? There's no cosmic teleology for technology to replace the things that make us human. We could reject the idea.

    It really comes down to the greed of a few. Witness the reasons behind the Screen Actors' Guild strike: the corporate studios want to pay someone for a single day's work to digitize their image and then use AI voices and computer generated animation to use their image as a virtual actor in perpetuity. There are so many ways that technology could enhance the acting profession, and make possible things that never were. But instead we have corporations investing millions of dollars in the goal of mechanizing human beings out of the equation entirely. And it saddens me to see so many cheer it on, with "You can't stand in the way of progress!"

    ReplyDelete
    Replies
    1. Greed but also a fundamental lack of respect for humanity itself, I think. We've lost the sense that there are things about being human that are good in themselves, not simply because of their instrumental value.

      I too get frustrated with the way the issue is portrayed - as though it is simply inevitable. It serves the interests of capital to perpetuate the notion that it's inevitable. But the wonderful thing about humans is that they can choose.

      Delete
  12. Your opposition strikes me as largely quixotic — as pointless as framebreaking. This technology will devour the field, will ye or nil ye. It's true that the wonderful thing about humans is that they can choose, but they *will* choose, and as always they'll choose cheap and abundant. It doesn't matter that two hundred years after Ned Ludd the best cloth in Britain is still woven by individual crofters on pedal looms, most cloth worn by most people will *never* be made that way, ever again. People just *do not care* about quality, and eventually we'll get to a place analogous to where we are now with fabric, where the vast majority of people can't even identify quality by either sight or feel, and balk categorically at the cost of it. (Notably, and in opposition to some other commenters, I also think the major RPG publishers are already doing something very akin to AI art production but using humans, and the art in the two latest editions of D&D, to take the most prominent example, has been absolutely soulless, dull and uninspiring, for all its high budget and technical lavishness; the pictorial equivalent of a Marvel film. As stimulating to the creativity as chewing cardboard.)

    What is do think is true, and where I think there's some hope of resistance, is that as Lum implies, making AI-generated character portraits and the like impoverishes play groups who might otherwise have drawn their own, and since this robs the game of some of the exact creativity that is one of its raisons d'ĂŞtre, I imagine many players will reject it. I think this logic also expands to our own little niche of the RPG hobby, where people produce labors of love for the joy of it and many insist on as high a quality as they can possibly obtain — sometimes even to the lamentation of certain consumers, who don't like the heavy focus on full-color, Smyth-sewn $50-70 hardcovers. Here we can perhaps expect to continue to see good work done well, illustrated by competent artists, or in a pinch those public-domain Victorian illustrators who are still so hard to match, let alone outdo.

    ReplyDelete
    Replies
    1. I don't think it's quixotic at all. If I convince one person, then that's one person more than there would otherwise be who agrees with me, and - probably more importantly - we need to keep the flame of humanity alive as we enter a period in which strong forces are seeking to extinguish it. We're in Alasdair MacIntyre's world, and I think the job of anybody who cares about these issues right now is to preserve what we can so that a rebuilding effort can take place once events have run their course.

      Delete
    2. Thank you for posting this and facilitating this conversation.

      Delete
  13. How mad are you going to be to find out I used AI to try to generate imagery for Yoon Suin? https://theamateurdungeoneers.blogspot.com/2023/08/i-used-satan-to-generate-illustrations.html

    ReplyDelete
    Replies
    1. Not mad at all, although I think the images reveal the shortcomings of the procedure. The results are serviceable, but kind of look 'what they are supposed to' look like.

      Delete
  14. I think the "margins" argument is the best one here - if there's some percentage of people that would have otherwise commissioned an artist using AI instead, it does mean some artists won't make rent when in a world without the AI art they would. I think even someone who is supports, or is at least okay with piracy, can understand the difference between piracy denying the already wealthy a few hundred thousand dollars in revenue (which is a generously high estimation, I think) and the difference between being able to eat or not. Marginal utility exists.

    As for argument 4 - "is AI Art art?" - I don't think it's really relevant. Someone using AI to generate a thing wants something of some passable level of quality to look like some description - whether it's "real art" or not is a philosophical concern that isn't present for them, at least not for the thing they're using the generator for at that moment. Basically, having an image generator that spits things out like a search engine and gives them images roughly in the shape they want is what most of these people are after, and concerns about it being art are not present.

    There are the actively deluded (or nihilistic) individuals that say they are "AI artists" and try to attach a price to things the machine spit out as if they put anything valuable into the process, but I think we shouldn't take those people seriously.

    ReplyDelete
  15. I broadly agree with you, but I do think that access to AI character portraits does make us [arguably] better off. Having a visual representation of your character often sparks off ideas about them that you would not otherwise have had. Would it be better if you'd thought about and drawn the character yourself? In many ways, yes, but there is also something to be said for the way in which those new perspectives arise from having somebody else do the portrait (with, in this case, "somebody else" being a gloopy mish-mash of every artist ever).

    I'm reminded of the Games Workshop/Citadel stalls at Games Days and Dragonmeets back in the 80s where, for £5, you could get a portrait of your character knocked out in 30 seconds by Tony Ackland (why, I used to ask myself, is it never John Blanche? In hindsight, it seems obvious why). As you stood there queueing for your portrait, you would see Tony knocking out stereotype after stereotype, with tiny variations. Other than the fact that these portraits were by a famous artist - a human one at that - the more varied possibilities offered by AI seem, to me, preferable.

    ReplyDelete
  16. Something else which is often missed when discussing AI and its effect on artists & writers - something which "problematises" your first assumption, but which I hope your "strict utilitarian" would pick up on - is that part of the issue here is money, and the need to earn it.

    Curtailing an artist's ability to make money from their art is only an issue in a world where their survival is dependent that ability. Part of the promise of technology was that it would free us humans up to do the things that we enjoy, and leave the tedious stuff to computers. Of course, that has spectacularly failed to happen, but we are approaching a singularity where something is going to have to be done with us increasingly unproductive fleshbots: either we're found new work that we can do cheaper/more efficiently tech can't do, we're all killed off, or we're given more leisure time - Universal Basic Income is one obvious attempt to steer things towards the third option.

    To put it another way: if an artist is only making "art" for commercial reasons, is it really art?

    ReplyDelete