People can argue back and forth about the degrees to which AI art is gonna take over the world or not do so, but this has always been the part of this equation that fails to line up for me from the evangelist crowd. Why is AI art worth pursuing beyond it being, like, “cool”?
I still don’t understand why this is something that is actually valuable to automate under any framework beyond monetary. If we were to progress technology along these lines, far enough that all labour is automated, people would be freed up to spend their free time doing whatsoever they please, and so much of what people please to do is to create. Why, then, is there inherent worth in the automation of artwork on the scale proposed in statements like this beyond “ooh, aah, artsy AI”?
Being technologically impressive, and it is very impressive, doesn’t necessarily translate to it being valuable on a scale larger than an individual scale of “user needs artwork and is not an artist / does not want to work with an artist / (the most charitable one) does not have time to produce the necessary artwork.” Maybe it’s that I’m a bleeding heart pretentious snob, but I feel like to automate creativity.is at odds with the entire reason art – not just visual art, but music, writing, dance, theatre – is such a mainstay in human culture.
The cynical part of me wonders how much of the appeal is just the street cred of being able to achieve artistry through code where many people draw an arbitrary line between “creative” disciplines and the more STEM-y types – a line which doesn’t need to be drawn, mind you. The optimist in me hopes that the push for AI art is just the prelude to proving AI can find use in more practical areas.
AI art will eventually reach a point where it can produce content of a similar or higher quality to humans at a significantly more efficient rate. People who consume art (music, literature, film) far outnumber the amount of people who produce it. Why should potential advances in the quality of art be held back just for the sake of the few?
I’m kinda low on time at the moment, so I’ll be back later, but:
Because when you create art with the intent of appealing to a wider audience, there is a basic standard of quality. AI allows one person to create so much more without sacrificing their creative vision to a group of others, each with their own vision, that can still reach others. AI is by no means necessary to achieve this goal, but it sure helps.
I don’t think that’s what they were saying. I think it’s more of a “the creation and consumption of art is beneficial to humanity”, which is a sentiment I wholeheartedly agree with, and “AI interferes with these goals”, which I think is the primary point of contention here.
What I find funny with that assertion is that it implicitly views itself as the only valid viewpoint.
A lot of people dislike creative endeavors but enjoy the simple, rhythmic nature of “dumb” manual labor. They’ll happily toy with AI art for fun and without a care for the artists living in their bubbles that do not intersect with theirs while lamenting the rampant automation.
I guess that because they feel this way we should roll back all of the industrial revolution and dedicate automation efforts solely on art instead ?
Art gets automated because everything is getting automated. This is what ‘progress’ as loosely defined tends toward: efficiency in every domain. Art isn’t some sort of special unicorn facing a new and unknown threat in that regard, artists are merely the latest target of a long line of innovations leading from the XIXth century to here.
I guess I’ll chip into this, since it’s part of why I shifted myself towards limiting the stuff I post here significantly—
I despise the current wave of AI taking over various sectors. Automation in cars terrifies me. The algorithms, in spite of how stupid they are, are way too quick to adapt to even the words we say around our phones to show us ads we’d want to see.
And the AI art craze is no different.
There will be no replacement for artists, as I think everyone here has said so far: AI art can’t even come close to replicating the intricacies that the last decade of spriters hasn’t been able to already reach. If we look at Klok’s first post, we can practically see the crust on so many of those sprites.
This gives me so much despair. We all know that these algorithms are not going to listen to what we want. There is not a soul innocent enough who uses or creates AI that won’t have their tool accidentally or purposefully take art pieces that aren’t ‘free to train on,’ or that individuals won’t take the code and personally train their version of the AI off non-F2T art.
I won’t talk down on Klok’s post in spite of how much I disagree with it, but the statement of “Do we now what sources X was trained on?” are irrelevant because the AI doesn’t simply learn from each piece it views, but instead can mirror it in a way that Human natural error couldn’t reach. We’re simply lucky that AI has been much, much, much slower on picking up the pace, doesn’t understand the human body nearly as well as humans and can’t truly replicate human creativity.
I’m not a good artist, but everything of mine would be permanently labeled as not F2T, including my maps and battle animations, in the event people figure out how to get those working with AI tools.
There’s no ethical way to create an AI that’s open source which has no avenue to abuse it to steal art. The only ‘ethical’ way is for a trusted community member to make an AI trained solely on art submitted to be trained on, as well as t having a ban of every vanilla GBA FE sprite. The minute anyone else gets access to its code and is able to make their own version, we will be rife with abuse.
Right now, we are, as I think several have alluded here, at a situation not unlike the work force right as industrialization or modern computers and automation began to make the human less important in fields like item creation. However, we are not looking at menial tasks which were dangerous for the individual, but now we’re threatening the ‘creative freedoms’ that people were allowed to take when they weren’t dying from working in slave-like conditions in factories or barely eking out a living as a sustenance farmer. We prided ourselves on the fact that people were now allowed to become more creative with their free time and express themselves more vigorously, and we’re now choosing to further limit the ways we’re able to express ourselves and earn money while doing so.
Perhaps if the proponents of AI were so wholehearted and charitable to do this simply as a way to advance the way we think about art, I would agree with you. But none of them are.
I think Zeldacrafter’s post is excellent here and really summarizes my points more concisely than I ever could, but I want to make another thing clear: I’d rather waste several hours of my life working on a splice for someone for free than see them generate a soulless portrait in 3 seconds by shoving a prompt into a machine. And I heavily dislike doing that for any reason.
I just want this all to stop so we can get back to having fun and doing art without any threat in any shape to our work, man.
It may be the idealist speaking, but the pourpose of automatisation in our civilisation is to achieve (at least partial) post-scarcity. When we automatise everything, there would be no reason to exchange our labour for monetary gains. This, quite ironically, would be artist paradise. Imgaine if you just could make art because you want and not because you need it to put food on your table.
This is in future, but that where automatisation (AI included) is leading us.
As someone who’s been learning how to draw for around two years now and still having a long way to go, I see AI generated art as the devil. It’s like cheating in a (chess) game. Instead of learning all the ins and outs (shading, anatomy, character design, colour, etc.) AI allows you to just magically pull a bunny out of a top hat. It’s unfair to those who have been trying to learn a craft. Kinda sad even, that I see people online get angry abou “their” AI artwork getting “stolen” by other people.
Funniest of it all might be that I’m an IT student. I should be blown away by the tech behind it right? Partially I am yes. But what overshadows all of that, is the sheer amount of problems AI art and AI in general brings to the table.
AI Art kills creativity and steals from individuals that have practiced a large portion of their life to achieve a level of skill. I can definitely see the benifits of AI Art, but I cannot fully understand why some people on the internet defend AI Art so vividly even if it is cool tech. Would be quite the bummer if all of a sudden a job you studied all your life for gets stripped away by a tool that allows anyone to do what you’ve been doing for years.
As someone who’s going to spend possibly the rest of their entire live diving into art and its techniques and principles, I support the voice of artists about the problems of AI Art. It needs clear rules and needs to be a distinguishable seperate field to the standard craft and the Art theft needs to stop.
Think I’m just gonna reply to everything in a big fat wall and head out, since I think there’s a pretty hard philosophical barrier here that’s hard to breach and I don’t want to be raining on people’s parades in that case.
Response
I ask again, “what will people do if all practical tasks are automated?” If we don’t need farmers, accountants, lawyers, truck drivers, etc… The logical endpoint of trying to argue that this industry should be automated is that all industries should be automated, no? So we eventually progress far enough that this is where we are technologically; what are people now doing with the ubi you propose we’ll start receiving and the massive amount of free time? Just sitting around consuming AI generated algorithmic media until we die? I’m not asking why we’re trying to apply AI to the field of art, I’m asking why there are people who are of the opinion that automation of art > nonautomated art.
Not all people are artists by trade, or are particularly skilled at it, but I believe people are naturally drawn towards trying to express themselves. Art isn’t the only way we do it, but it’s one of the most enduring, which is why we’ve seen evidence of our ancestors creating art from our very earliest days. There’s a weird tendency in these discussions to act like art is some thing that only the creatively gifted partake in, as though there’s an objectivity to the quality of art that imbues it with more worth. I’m sure AI will one day be able to make the most photorealistic black and white Walter White pencil drawings ever posted on reddit, but that doesn’t mean that they’ll be creating “better” art than people.
It’s such a VERY bold claim in the first place to claim that AI will produce higher quality art than a person can. Bold enough, even, that I’m confident in saying that it’s entirely unfounded. Even if AI can reach competence levels equal to or higher than the average skilled human artist, AI don’t actually think, despite all the anthropomorphism that gets tossed around in these kinds of discussions. They don’t have lived experiences, they don’t form or hold opinions, they don’t understand meaning. We’re not programming Cortana, we’re programming very advanced algorithms. During a point in time where people are the most up in arms about design by committee corporate artwork that they’ve ever been, why are we trying to argue that an unthinking pattern recognition robot is going to be the future of media?
If your argument is that AI will be really good for churning out, like, the kind of low impact artwork you might see on a corporation’s website or in a brochure or something that’s purely based on aesthetics, then sure, I’ll buy that – there are people who are probably very passionate about creating these things, but they’re definitely labour first and foremost and automating them makes sense. Same for things like technical manuals or other similarly practical applications for AI generated writing.
But when the discussion starts getting into this territory of “one day the AI will spit out art that we will consume, as is our natural inclination,” I start to lose the plot heavily.
This is such an interesting framing every time, as it assumes that the argument is some kind of “technological progress vs traditionalist luddites” framing where the artists are just afraid of innovation. There is more art than it would ever be possible to consume in a single lifetime being produced every single year, and I still haven’t been given an answer as to why it’s important or relevant that there are machines that can produce more media, faster, that isn’t from the perspective of “it will be cheaper.”
This is part of why I said this; there’s some strange phenomenon going on, here, where machine is being put on a pedestal for reasons that escape me. It’s also why I use the term tech evangelist; A machine is not inherently better than a person at creating the things they are designed to do just because they are more efficient or do so more “perfectly,” especially once you start trying to apply these things to fields like art, where things are so so subjective and so so so rooted in trying to communicate something to an audience that the author truly believes in. Telling an AI “write a medieval story about the value of self-worth” or something similar is not actually equivalent to having somebody with an opinion on the matter write such a thing, even if the AI will spit something out faster and have it be algorithmically perfect, because the AI doesn’t believe anything it’s outputting. It doesn’t believe anything. It only knows what combination of phrases, imagery, plot, and character will create a facsimile of belief.
Response
I don’t think it interferes with the creation of art on its own at all, which is why I didn’t come into the thread earlier when the discussion was just about the application of AI art in a general sense. It’s specifically the endpoint being argued in statements like the one I responded to, where one day art will entirely done by a machine, that confuses me.
Partial automation of the creative process? Great. We already have tools that do stuff along these lines, so making them more advanced sounds good to me.
Automation as a tool to help an artist create artwork they’re happier with? Sure, it’s a tool to be used in the creation of art, so I’d sure hope it helps artists somehow.
Automation as some kind of logical endpoint to everything, where one day AI art will be both superior and numerous, and that this is something that is not only achievable but desirable? I understand that the contention here is primarily philosophical, but that’s where I really disagree.
Response
This is mostly true, but surely I’m not the only one who sees the difference between, say, a fully automated welding job and a fully automated movie production, right? While they might not be substantially different experientially while they’re being produced, as both are just work that produces an output, the way people “consume” a good bit of welding isn’t actually equivalent to how we “consume” a good book, or an album.
The “Surely if you oppose thing A, you will also oppose ostensibly similar thing B as well, right?” school of argumentation is pretty lame, don’t you think?
Response
Sure. Like I said before, the way I see it is that the value of a fully automated workforce is that it would free people up to spend their time doing the things people like doing, and a large part of that is the creation and consumption of art – and, again, throwing back a quick reminder that I think a far larger portion of the populace would partake in casual creation if there weren’t monetary pressures on everyone to try to monetize their interests and thus only pursue that which they’re most competent in.
I think the reason I don’t 100% buy this framing, ultimately, is the way a lot of people I’ve seen advocating for AI art advocate for the scale of it. If people are 100% free to do anything they want with their time, but we have AI spitting out hundreds of thousands of pieces of media every day, it becomes so much harder to actually consume media as created by people, as it will just get drowned out. I think part of what makes media valuable is that it’s a way to communicate these complex ideas, and then develop on those ideas even further through the discussion of that media after its creation. This is why, again, I really don’t take issue with AI being used in artistic fields, but I do take issue with the idea that they should be used as a replacement for an artist. There are surely ways we can, as a society, decouple the creation of art from the need to make money that doesn’t also involve such consequences.
There’s then, of course, an absurd number of philosophical quibbles you could get into about the nature of who chooses what kind of art the AI produces and the effect this would have on human society and etc, but like I said way at the top of this whole screed, I don’t think I’m interested in trying to break through any more philosophical disagreements than the one I’ve already planted my flag into.
Although, as an aside, if the main way AI art would be valuable would be once we reach that theoretical end goal, is pursuing this status quo now instead of then valuable?
Response
This is part of what’s interesting to me here, bc a huge portion of why people are getting into the AI business right now is not for some utopian endgoal, but for money. A lot of ink is spilled trying to assuage those worried about their livelihoods that AI art is not a threat to their income and that real artists are never going to truly go away, while simultaneously saying that one day AI art will, along with all other innovations, permit people to not have to rely on art as a source of income. These aren’t compatible statements; AI art either will or will not replace the field of art as a profitable industry to be a part of, and art is already very hard to make a lot of money off of in the first place.
If the goal is to fully automate the industry, then yes, artists are right to be anxious about their futures, so arguing that this is both the ultimate and inevitable goal of this technology while also arguing that artists need to stop worrying about it is hypocritical. While in the absolute long term these are goals that are, ostensibly, worth pursuing and valuable, we are currently in a system that will use these things not as a way to replace human labour with absolute freedom, but as a way to increase profits.
Until this hypothetical post scarcity fully automated society comes to fruition and there’s no longer a need to work to live, it’s irresponsible to ignore that that need still exists, especially if you’re one of the people poised to actually profit off of the AI field yourself. It’s not difficult to see why a lot of artists view this hypocrisy as being indicative of a lack of respect for their field – an assertion that not being able to support yourself off of art is, ultimately, not that big a deal, but that the field of AI research is inherently more noble and deserving.
There must be an acknowledgement that you can pursue technological progress without necessarily trampling those who don’t stand to benefit from that progress underfoot. The outcry isn’t a cry to halt that progress, but to progress compassionately.
Yeah, fair enough. I still am not convinced whatsoever, though; I can acknowledge that you can advance technological progress without trampling those underfoot, but I also have 0 faith given how little I’ve seen of so, so many who are so willing to. I’m sure I’ll find someone who makes a convincing argument or truly makes genuine progress, but that has NOT happened yet.
I had always assumed that AI art would be used to assist a human in communicating a vision ; that there’d always be a human directing the tool and using it to create something - and we can have plenty of debates on whether it’s a good or bad thing to give one human the power to create a movie from scratch with AI actors and set designs and all that jazz, in just a few clicks, but ultimately that’s sorta irrelevant and elitist anyway.
What Xilirite seems to suggest is a future where AI isn’t a tool but the end goal, and an algorithm similar to the one YouTube uses for recommendation would just up and create ‘customized’ ‘art pieces’ from scratch.
That’s, uh, something I hadn’t even considered. And it does sound pretty terrible when we see the dire effects of botting on social medias and search engine right now - if the algorithm wants man-made content buried it’ll succeed.
Alright, a lot has happened in this thread today, huh? Welp, best get started.
Reply
I mean, there’s tools out there that can tell if a paper was written by AI, I’d imagine a number of anti-AI abuse tools are in the works. I do agree that trusting strangers on the internet with a machine that is capable but not intelligent (a bad mixture in people, worse in machines) is dangerous and will require constant vigil, but there are two reasons we must keep that vigil instead of attacking the AI:
There’s nothing we can do about AI as a whole, only about specific abuses of it. Therefore, we have a moral duty to rectify those abuses, and to mitigate evil where we find it.
The AI, so long as it’s guided by a human, can and will have positive effects on the quality of art. I know this is the point where we’ll disagree, but the fact that some human artists are indistinguishable from AI says one of two things: either that many human artists are as devoid of understanding as the AI (my personal opinion), or that the human element of art is completely irrelevant, which is something I don’t think either of us believe.
Reply
This may be the cynic speaking, but we’ll either destroy ourselves before we get there, or the oligarchs will keep the prosperity to themselves (and please no one bring up overthrowing the oligarchs- every society has them, no matter its nominal structure, and there’s no changing that, only who the oligarchs are). Also, even assuming that happens, I don’t think the complete automization of art would be in any way beneficial, for (as I’ve previously stated) I think there is a human element inherent in good art that an unthinking, emotionless machine will never fully match, only ape. For other fields mechanization is a great idea, but for art not. Trading out the singular cook and gourmet recipes for a team of summer camp food workers and their instant eggs is great to feed a starving populace, but when it’s quality that really matters, robots and AI can’t match the greats in the realm of art.
Reply
This doesn’t sound like you care much about the soul of art, it just sounds bitter. By saying AI art allows non-artists to cheat, you’re conceding that AI art is as good or better than human art. If that truly is the case, than there is no reason to oppose AI art other than spite. In fact, that’s the vibe I’m getting from most people complaining about AI art.
Reply
Sorry, yeah, this miscommunication was mostly my fault, I was running short on time. I didn’t mean to imply that you thought all implementations of AI harmful so much as cutting the human element out was destructive of the very ends of human art. I think we mostly agree here.
Reply
Okay, this one pissed me off. Perhaps unreasonably so. I’m going to try to keep it impersonal, but this whole idea makes me real mad.
That’s a subphilosophy of the “Death of the Author” philosophy, which is one of the stupidest things to ever come out of postmodernism. Intent maters. It’s not everything, after all, if I say there’s a road paved with good intentions you know which road I mean, but it does matter, and any attempt to say it doesn’t is postmodernist subjectivist nonsense. The fact that AI has no inherent intent is a large part of what makes it inferior to the human greats. If you don’t believe in objective truth that argument might hold some water, but if you do, like me, than it’s taking on water instead.
Alright, that seems to be about it so far, and though I’m not a mod or anything, I’d like to thank everyone for maintaining mostly civil discussion (even if I’m not always the perfect example of that).
If this sentence is in reply to what I said, I’d like to say that I didn’t say intent doesn’t matter.
The fact that AI has no inherent intent is a large part of what makes it inferior to the human greats.
Maybe it doesn’t right now. This is more in reply to a sentiment I’ve seen in this thread that AI will never have some quality that would be inherent to humans, but I don’t see how any of our qualities can’t be acquired by AI. Maybe it’s not possible right now, but one can’t write off the future.
If you don’t believe in objective truth…
What is objective truth? I’m also unfamiliar with post-modernism.
My only concern about Ai generated Images, is the harm it will do to potential new artists.
I started spriting because i wanted cool new characters for my dumb hack, thanks to that spriting is a new hobby of mine, i have been trying to be a little bit better every now and then for a good while.
If i could have just generated an impressive new sprite from an Ai prompt, would i have spent these 2 years practicing? I highly doubt it.
Of course, there are cases and cases, but thats how I see it.
I’m sorry if I did not put my thoughts into words correctly, but there seems to be a misunderstanding here. The reason I do not support AI art is not only because I think AI Art produces “better” results than human artists. There is also theft involved in the process of “training” an AI.
I appear bitter yes, but I don’t see why that is a problem? Like I explained, I see AI Artwork generators as cheaters because the program or algorithm never had to learn the theory behind art. There are skills needed to become an artist and everyone can learn those given enough time. I am “bitter”, because something I put effort in learning suddenly becomes outmatched by something that never had to go through so much blood sweat and tears. The fact that everyone using a tool can skip a learning progress is unfair in my eyes. Imagine being a professional sportsman or sportswoman. You work years on end to get a perfect condition and learn the ins and outs of your sport. Suddenly, a wonderous new sports drink gets released that allows everyone to get your level instantly. Of course you’d feel negatively, knowing that others get to skip the entire process you worked so hard for. They might even surpass you as a result.
Now, I’ve been learning how to draw Manga for around two years now and there are people who have spent way more on it. What’s still unfair in that regard, is that the AI is just much faster at producing art or something in general. Regular artists will get drowned out by others mass generating images and that is another reason why I think that there should be better rules for AI Art as it stands. Again, you don’t want to get replaced by something that does something better or faster than you.
As for the quality of the artwork, I think that AI Artwork can be decent at times and it will undoubtedly become way better in the future. If everyone can do it, what is the purpose of being/becoming an artist then? Art gets its value because there is a person with emotion, skill and passion behind it. Not a robot that pumps out works for people to admire.
This is me as well, hence why I agree with this so much. Why start if others can do it better instantly?
Like I said, I am not completely against AI Art. It can actually be a very useful thing for quickly visualising thoughts and ideas. It can even be used to learn techniques or styles if you do not want to only study the theory behind it all. There just need to be clear rules on the subject. No more Art theft from Artists for an example to train these image generators with. Websites should (in my eyes) seperate AI Art and Human Art, so that both forms of the craft can be respected and admired. But as it stands now, things are going the wrong way if AI improves and things remain as they are now.
I hope I do not sound toxic, because that is not the tone I’m trying to achieve with this reply. Last thing I want to do is gatekeep and present myself as someone who is entitled to their opinion.
AI is here to stay. Like it or hate it, there’s nothing you can do about it. A Pandora’s Box, once opened, cannot be closed.
As an artist, you can do two things:
a) ignore the fact AI exists at all and continue as if nothing has happened.
b) learn to utilise this new technology
There’s no option c). You can try boycotting it now but such a boycott won’t last.
That’s an excellent question, but the answer is simple.
Because you want to.
Why learn to drive if you can take a tram? Why learn to write if you commission someone?
Maybe because you want to have a skill someone else doesn’t. Maybe I’m not seeing something, but I can’t understand why someone would stop drawing just because AI generation exists.
Not replying to anyone in particular, but just wanted to get my opinon out there.
I recently watched a very well made video about the history of synthetic voices, and there impact on the voice acting industry. The video ends with asking the question of if Ai voices will replace human voice actors, considering that at there best, 99% of people can’t distinguish them from real voices. To this question, the creator responded by saying something that I think can be applied to this issue very well. It went somthing like this (paraphrased because I don’t remember it word for word):
“How many people generating free ai art (portraits or otherwise) would have been willing to spend the money it would have taken to get that art commissioned by an actual artist (or the 100s of hours it would take to learn how to make it yourself)? How many people who use stock music in there videos would have gone through this same effort? These are alternatives to NOTHING, not to professionals.”
(I’ve added some notes in brackets like this to create a analogy between these two topics)
By allowing ai portraits, you aren’t taking anything away from real artists. Your only empowering people who wouldn’t otherwise be willing or able to explore these passions. People who really care about the human element and think its essential won’t instantly jump ship if ai portraits are allowed so human artists will always have there place.
What I’m trying to get at, really, is that I don’t think human and ai art have to compete with each other. As long as they remain separate (perhaps in there own section of the repo, or there own repo altogether), and monitored by trusted individuals (not just some guy on 4chan). I simply see no reason why ai art should be considered dangerous or threatening.
The “Surely if you oppose thing A, you will also oppose ostensibly similar thing B as well, right?” school of argumentation is pretty lame, don’t you think?