|
Post by dragonaltar on May 22, 2024 12:57:22 GMT -5
Are you speaking of everyone who is continuing to use AI tools despite the opposition?
Yes, if someone becomes aware that the majority of visual artists are vocally opposed to it but decides to do it anyway, it comes off as a statement that the way artists perceive the situation doesn't really matter to them. They continue with what is essentially a fun little novelty at the perceived detriment of artists. To us, it comes off as saying "I've heard you, and I don't care." "Perceive" and "perceived" here is key as it doesn't require an agreement between the art crowd and the AI generation crowd on The Issues, but rather a lack of respect–generally, if a group of people says "hey this thing could be really harmful to us," it's best practice to stop even if you don't think the harm is major.
Some notes and concessions I'd like to make as someone who does illustration, abstract, and glitch art:
"It's ugly/soulless": I'll accept soulless as a vague concept, because it damn sure feels like it, but ugly is subjective and not terribly helpful. I make a lot of ugly art, sometimes on purpose. I will say that without the understanding of an artist, finer details that bring out "the magic" tend to get washed away in generated content. It's just averaging the patterns and shapes that it's asked to generate—often poorly. It won't have a unique identity. As someone who prefers striking work that bares the soul of an artist over technically impressive or realistic work, it has little value to me.
"It is lazy/low-effort": A lot of good traditional and digital art is made fast and with little effort. Also not terribly helpful as an actual argument.
"There is minimal human involvement / the person sharing it did not create it": I'm not against procedural generation and computer-generated art as a concept. I personally don't like this iteration of it for reasons I stated earlier.
"It is impossible for the prompter to connect with AI art in a meaningful way": People can connect to anything. Maybe they just like the colors, and maybe that's enough. But I agree that you will never experience the very visceral sensation that part of the artist is in there with prompted imagery.
"It breaks immersion in the fantasy of the past because it is such modern futuristic tech": I understand this viewpoint but disagree, the way in which something is created is generally a meta element, and peeking behind the curtain to find out how something was made will almost always break immersion.
"Some AI enthusiasts like to troll/antagonize those who oppose AI art": I brought this forward primarily to explain why artists have come to feel so strongly about it, and will also note that more than just "some" AI enthusiasts do this but not always on purpose—sometimes it's obviously very deliberate but often they just say very naive things about art and artists, which is frustrating and strengthens bad associations.
"It is hurting/will hurt the profits of commercial artists and maybe even put them out of work": Not my area of expertise but we do see companies using AI work that people used to do.
"It learns from copyright works without permission / is stealing": This one's complicated. I don't like copyright, as it's usually used as a bludgeon by major corporations instead of used to protect smaller artists. I don't care when people make traditional collage art. But I also don't like seeing small artists' works being used to generate art slurry.
The only thing I noticed you missing from your list was regarding the excessive amounts of water and energy it takes to keep the data centers dedicated to AI up and running.
|
|
|
Post by gambali on May 22, 2024 19:28:29 GMT -5
May 20, 2024 20:58:28 GMT -5 magicwillow said: P.S. Above statement is completely ignoring the AI debate.
Very true. AI seems to be a promising development but as far as the creative arts it's not something I'm terribly interested in, and therefore not a debate in which I'm qualified to participate. I do appreciate the different viewpoints here. A learning moment!
RE: DS orthodoxy: I hear a lot of different things in DS. Black metal and 90's computer game for sure. But also Berlin school, prog rock, CMI/World Serpent post-industrial, the hauntology of bands like Broadcast, neo-classical, minimalism etc. etc.
Cool to see people here taking things in new directions.
May 20, 2024 21:29:28 GMT -5 moreteeth said: I listened to your Bandcamp. Has a nice Goblin vibe to it, as in the Italian rock band.
That's a huge compliment, thank you.
|
|
|
Post by lolth on May 23, 2024 9:07:37 GMT -5
I wholeheartedly agree. Even if you were to feed an algorithm the same stuff that usually constitutes DS album covers (copyright free D&D, medieval, fantasy, etc, art as mentioned before) you're still presented with a problem. When you find an image and decide to feature it as your cover art, there's a process of experiencing the piece and feeling the connection between it and your music. It's not a pure aesthetic consideration either; you develop an understanding of the image and its context and go through the process of deciding if it's right for the purpose. I would hope the process of finding/making cover art is a little more involved than 'this is a random thing that looks fun', anyways. Especially for a genre like DS. It's also important to acknowledge that the temporal immersion is almost immediately evaporated once you realize the art used is a product of artificial intelligence, and that goes for almost all media across genres. It's like a pair of converse shoes in the background of a film about Marie Antoinette. When a genre is a child of eighties video games and ambient, using artificial intelligence shouldn't even be a notion if you have any consideration for craftsmanship. On a personal level, I'd never understand why plugging some prompts into a box and watching sludge come out is more rewarding than finding the perfect image, editing it, putting it through a Xerox, and ostensibly creating an aesthetically viable product using the means available to you. I'll elaborate upon this later. Very well said. Dungeon synth has an old-school, DIY work ethic. I don't care if your style is a futuristic high-tech sci-fi style, I still believe there should be some real human input in every part of the process.
|
|
|
Post by andrewwerdna on May 23, 2024 10:12:39 GMT -5
The only thing I noticed you missing from your list was regarding the excessive amounts of water and energy it takes to keep the data centers dedicated to AI up and running. Oh right, thank you. My response to that is pretty straightforward. I think it is with AI that we have the best chance of solving the climate crisis (some evidence that it's already doing good: www.weforum.org/agenda/2024/02/ai-combat-climate-change). And when it comes to AI art generation, that's like the energy use of running a graphics-intensive video game, so the same criticism could be leveled at gaming, just that it's more spread out. I don't know if you've noticed, but anti-AI people say a lot of cruel things as well. I don't do much social media beyond lurking the Facebook DS groups, but from my experience it seems that the vitriol is primarily coming from the anti-AI side. I'm sure pro-AI folks lash out quite often as well, I just have not personally seen it. Are these things actually relevant to the issue at hand? What we're grappling with, this question of AI acceptance, it's much bigger than these humdrum day-to-day social media spats. I understand the feelings, but I think the conclusions they lead to are short-sighted and missing the forest for the trees. I think we may need to pass through this forest. I'm not arguing that we should all rush in headlong, but merely that some of us should be allowed to scout ahead and report back with our findings. If there's anything I'm sure of, it's that we don't know what this is yet. Anyone who acts like they know is either ignorant or bullshitting.
I think when used well AI art has an alien soul. Maybe it can never contain the full soul of the artist/prompter, as much as traditional art at least, but I think at times a sense of the presence of a different soul emerges, or perhaps like an amorphous soul-stuff. I think this will be more apparent with AI music than with AI visual art. I personally am totally fascinated by these possible alien/ghost/blob souls, but I can understand how that may not interest everyone, and may even scare some people. But that is just the nature of technological progress. Should we have never started doing photography because portrait painters were vocally opposed to it? That's not to say every new technology should be implemented without considering the consequences, but I think people are too quick to jump to positions of certainty about what it will entail. What I'm really bristling at is the idea that everyone who uses a technology is in the wrong simply because a group of people believe that it might end up bad. By that logic, if AI art turns out to be a boon for humanity overall, then you are doing harm by holding it back. I believe people generally want to do good, but we are just each working with very limited and diverse data sets. Also it is frustrating that you do not differentiate between those who are trying to work these tools in a way that is as nondisruptive as possible and those who use them in a flagrantly antagonistic way. Apparently just everyone who uses them is guilty, and there is no way to ever use them properly. I think that is not a reasonable take. You might say that it is just too early to start using these tools, but I think the only way we can get a clear perspective is for at least some of us to start working with them in earnest to see what they're really capable of, and appreciating them for what they are without social distortion. Perhaps it really is just a bunch of derivative soulless detritus (in which case we have nothing to worry about), but perhaps it is something more. If we shut down all balanced discussion and are intolerant of anyone sharing their AI art in these spaces then we will not find out until it becomes normalized in larger society and slowly seeps in, forcing us to adapt to their framework of how to use these tools. Or we could see this as an opportunity to get there first and shape the foundations of the future. I do not think it is just "a fun little novelty" (though it can be in many situations). I think it is profoundly powerful beyond what we can imagine, and I think that this is actually a criticism that could be added to the list, that it is an existential Lovecraftian fear of the unknown, a yawning alien abyss of infinite aesthetic potential. Imagine what this tech might lead to in 10 years, 50 years, 100 years. I think we will each be able to construct our own personal fully-immersive procedurally-generated dreamscapes, teeming with all sorts of beings that seem every bit as convincingly real as biological creatures. This is the vision I want to work toward. It is awesome and terrifying, inspiring me intensely and making me very excited to do creative work right now. So yeah, not just a "fun little novelty," but rather something beyond important for art and humanity in the future I think. And of course there is room for hearing concerns, discussing them in depth, and having a difference of opinion. I'm not shrugging off the concerns, it's just a lot to grapple with all at once.
|
|
|
Post by moreteeth on May 23, 2024 14:24:27 GMT -5
The only thing I noticed you missing from your list was regarding the excessive amounts of water and energy it takes to keep the data centers dedicated to AI up and running. Oh right, thank you. My response to that is pretty straightforward. I think it is with AI that we have the best chance of solving the climate crisis (some evidence that it's already doing good: www.weforum.org/agenda/2024/02/ai-combat-climate-change). And when it comes to AI art generation, that's like the energy use of running a graphics-intensive video game, so the same criticism could be leveled at gaming, just that it's more spread out. I don't know if you've noticed, but anti-AI people say a lot of cruel things as well. I don't do much social media beyond lurking the Facebook DS groups, but from my experience it seems that the vitriol is primarily coming from the anti-AI side. I'm sure pro-AI folks lash out quite often as well, I just have not personally seen it. Are these things actually relevant to the issue at hand? What we're grappling with, this question of AI acceptance, it's much bigger than these humdrum day-to-day social media spats. I understand the feelings, but I think the conclusions they lead to are short-sighted and missing the forest for the trees. I think we may need to pass through this forest. I'm not arguing that we should all rush in headlong, but merely that some of us should be allowed to scout ahead and report back with our findings. If there's anything I'm sure of, it's that we don't know what this is yet. Anyone who acts like they know is either ignorant or bullshitting.
I think when used well AI art has an alien soul. Maybe it can never contain the full soul of the artist/prompter, as much as traditional art at least, but I think at times a sense of the presence of a different soul emerges, or perhaps like an amorphous soul-stuff. I think this will be more apparent with AI music than with AI visual art. I personally am totally fascinated by these possible alien/ghost/blob souls, but I can understand how that may not interest everyone, and may even scare some people. But that is just the nature of technological progress. Should we have never started doing photography because portrait painters were vocally opposed to it? That's not to say every new technology should be implemented without considering the consequences, but I think people are too quick to jump to positions of certainty about what it will entail. What I'm really bristling at is the idea that everyone who uses a technology is in the wrong simply because a group of people believe that it might end up bad. By that logic, if AI art turns out to be a boon for humanity overall, then you are doing harm by holding it back. I believe people generally want to do good, but we are just each working with very limited and diverse data sets. Also it is frustrating that you do not differentiate between those who are trying to work these tools in a way that is as nondisruptive as possible and those who use them in a flagrantly antagonistic way. Apparently just everyone who uses them is guilty, and there is no way to ever use them properly. I think that is not a reasonable take. You might say that it is just too early to start using these tools, but I think the only way we can get a clear perspective is for at least some of us to start working with them in earnest to see what they're really capable of, and appreciating them for what they are without social distortion. Perhaps it really is just a bunch of derivative soulless detritus (in which case we have nothing to worry about), but perhaps it is something more. If we shut down all balanced discussion and are intolerant of anyone sharing their AI art in these spaces then we will not find out until it becomes normalized in larger society and slowly seeps in, forcing us to adapt to their framework of how to use these tools. Or we could see this as an opportunity to get there first and shape the foundations of the future. I do not think it is just "a fun little novelty" (though it can be in many situations). I think it is profoundly powerful beyond what we can imagine, and I think that this is actually a criticism that could be added to the list, that it is an existential Lovecraftian fear of the unknown, a yawning alien abyss of infinite aesthetic potential. Imagine what this tech might lead to in 10 years, 50 years, 100 years. I think we will each be able to construct our own personal fully-immersive procedurally-generated dreamscapes, teeming with all sorts of beings that seem every bit as convincingly real as biological creatures. This is the vision I want to work toward. It is awesome and terrifying, inspiring me intensely and making me very excited to do creative work right now. So yeah, not just a "fun little novelty," but rather something beyond important for art and humanity in the future I think. And of course there is room for hearing concerns, discussing them in depth, and having a difference of opinion. I'm not shrugging off the concerns, it's just a lot to grapple with all at once. I think most people here only have an issue with artificial intelligence in so far as its generative and marketed as an art product, e.g., ChatGPT and Dalle. If it can solve the climate crisis, that's great; it'd be nice if artificial intelligence could solve all of our problems so we can sit around and make art all day, instead of doing both. I think that drawing a connection between photography and artificial intelligence is a false equivalency for more reasons than one. The camera was a revolutionary device because it was an entirely new art form, and that conversely changed how classical art was made; movements which highlighted the craft of painting itself—impressionism, for example—were created because the previous functions of painting as an accurate replication of reality had been supplanted. In essence, photography was an entirely different thing that served a similar function. However, artificial intelligence as its currently used is entirely derivative. So—just like digital art—the function of artificial intelligence is used to create art or at least something with the same pneuma as traditional art, and not a product of an entirely new form. In what way would artificial intelligence motivate traditional art to change in a positive direction, what does it do better, and what does it tangibly change? While photography replaced art's menial function and motivated art to differentiate itself via its craft, artificial intelligence imitates both its menial function and its craft while being considerably primitive at both. Art created by artificial intelligence has no advantage or disadvantage over normal art besides nebulous conceptions of availability, but its really just as accessible as a pen and pencil or a copy of GIMP. People just don't want to put a little effort in. I think that considering the people and community behind artificial art is a common feature of criticism against all art. The average artist could easily say that early digital and internet art was ugly, but its distinctive features were indicative of an era of experimentation amongst unorthodox pioneers, which adds what people like to call "charm" (see: Yvette's Bridal Formal or Superbad). Artificial art, as it is now, is primarily associated with tech bros and intellectually bankrupt corporate investors, which just makes the mushy, oversaturated style seem egregious. If artificial intelligence grows out of the state it's in now and can create art that's indistinguishable from traditional art, then I don't think anybody's going to look back on the aesthetic of its early days with much nostalgic fondness like a modern web designer would Yvette's Bridal Formal or a primitivist would cave paintings. There'd never be a renaissance of poorly proportioned generative hell spawn creations, intentionally plastered over with that characteristic rubbery sheen for effect. I think it'd all be forgotten about. As for your last point, I'd only ever be interested in art as long as I and other people are making art. A holodeck dreamscape entirely generated by an artificial intelligence sounds closer to societal suicide than the utopia you make it out to be; more akin to Brave New World than Island. It really isn't the end goal of art to create the ultimate unchallenging, consumerist product with little to no human input behind it.
|
|
|
Post by crystallogic13 on May 24, 2024 2:32:21 GMT -5
I have written and deleted many answers because they came up unwillingly hostile to the usual small but bitching crowd against AI.
I am not a musician, nor artist, so i'm not invested in this, for clarification. But I sure as hell ENJOY all the great things in general AI is bringing, and also of course worrying for its implications to all of us now and in the future, and in many philosophical and practical ways. BUT while this is a great discussion, this thread is for AI used specifically in album art and the witchhunting-bitching-gatekeeping that is PLAGUING the fb pages one year now with pitchforks (makes you wonder...) and now here.. There are already many arguments and counterarguments, and I think the points are there for anyone who wants to see them.
And this whole thing could be solved with just one thing, a rule of "if artwork is AI, note it" and we could all go home, but NOOOOOOOOOOO, like some fanatic vegans (not the normal ones that don't bother you with preaching), there is a f circle of people with the same fuzzy arguments like "soulless" which is completely subjective, "effortless" who doesn't stand and "prompting" can become QUITE a whole world on its own, in a niche niche genre where people do everything out of their pockets asking them to pay sometimes 100 +++ (!!!!!!!!!!) euros/dollars for a cover art that could be WORSE than the dreaded AI...
As already said to these worried people, HAVE YOU ASKED that dead artist his art you're using for making music with your personal political/religious references for example? Would he be ok ? I wouldn't like personally a nazi scumbag making music and putting my artwork 200 years after I died but it happens anyway, so HOW AI is worse than this shit ? or all those MS Paint -win 3.11 era- covers that immediately just scream "stay away" ?
Anyway, as I said, the monster is already out, use it for good, ignore it, whatever, but please don't tell people how to do things and in DS especially where people are more experimental, relaxed and chilled out, please.
|
|
|
Post by magicwillow on May 24, 2024 9:01:27 GMT -5
With so many visual artists, myself included, expressing disdain for this kind of image generation pretty loudly (for varying reasons: stolen source material, ridiculous energy use, potential for impersonation, etc.), it feels like our concerns are being shrugged off when people do it anyway. Are you speaking of everyone who is continuing to use AI tools despite the opposition? Speaking for myself, I am doing my best to consider every point that I've seen raised, and I feel like I could address each pretty thoroughly, but I do not think it would be very productive to go into debate mode right now. It seems like people are just wanting to share their feelings about it more than actually hash it out. But for the sake of clarity, I think these were all the points raised against AI in this thread so far (please let me know if I missed any): - It's ugly/soulless
- It is lazy/low-effort
- There is minimal human involvement / the person sharing it did not create it
- It is impossible for the prompter to connect with AI art in a meaningful way
- It breaks immersion in the fantasy of the past because it is such modern futuristic tech
- Obnoxious tech bros are the most prominent people championing it, making it lame
- It is proceeding forward unabated despite the backlash
- Some AI enthusiasts like to troll/antagonize those who oppose AI art
- It is hurting/will hurt the profits of commercial artists and maybe even put them out of work
- It learns from copyright works without permission / is stealing
Is it just me or is point 5 a pretty good point?
|
|
|
Post by magicwillow on May 24, 2024 9:11:16 GMT -5
Are you speaking of everyone who is continuing to use AI tools despite the opposition?
"It learns from copyright works without permission / is stealing": This one's complicated. I don't like copyright, as it's usually used as a bludgeon by major corporations instead of used to protect smaller artists. I don't care when people make traditional collage art. But I also don't like seeing small artists' works being used to generate art slurry.
I actually kind of agree with this. I am all for remix, sampling etc. I think a computer "remixing" a human's artwork is worse to me than a human doing it. This is entirely based on how it feels to me, but hey we're talking about art so feelings are relevant. It feels slightly frightening in a science fiction post-human terror kind of way that a robot can remix someone's artwork with minimal input, sell it to someone else (because a lot of AI art is paid for with subscriptions etc.), and the artist not even know about it or get compensated. It feels fundamentally different to me because the process is automated by code, and I for one think it feels slightly scary that human considerations such as moral, ethical or legal considerations are either non-existent or being outsourced to machine-learning. This is not an argument, it is not me saying that all AI is bad or that no one should use it. But it is the feeling I get having looked into how it works more recently. This is why I'm going through the process of updating all my covers to be public domain, because it just makes me slightly uncomfortable.
|
|
|
Post by andrewwerdna on May 24, 2024 11:35:48 GMT -5
I think that drawing a connection between photography and artificial intelligence is a false equivalency for more reasons than one. The camera was a revolutionary device because it was an entirely new art form, and that conversely changed how classical art was made; movements which highlighted the craft of painting itself—impressionism, for example—were created because the previous functions of painting as an accurate replication of reality had been supplanted. In essence, photography was an entirely different thing that served a similar function. However, artificial intelligence as its currently used is entirely derivative. I was not trying to draw an equivalency between AI art and photography, just using it as an example of a disruptive new technology. Maybe the drum machine would be a better example? That was mostly derivative and bemoaned by people who thought it would cause harm to them, but we would not condemn everyone who used a drum machine. There are good and bad ways to use it like any tool. I do think AI art and photography have a lot of parallels though, but I actually think AI art is more revolutionary. The key phrase you said, "as it's currently used..." I admit I cannot think of a clearcut example where it has been used in a way that is totally novel, but I can imagine ways it might be eventually. For example, what if, when the models get efficient enough and the compute cheap enough, we could have "images" that are constantly being generated (like 24+ fps)? So let's say you had a landscape of a tree on a cliff near the sea during a storm, the wind could be blowing the grass and leaves, the tree swaying, the branches rustling, the clouds moving, the waves crashing against the shore, etc. Or let's say you had a photorealistic portrait of someone on a busy street in New York, the person could be smiling for the camera, sometimes laughing, looking around, shifting their position, scratching their nose, and all the while behind them there are cars and people constantly passing through the frame, never repeating. This might seem like no big deal, like just watching a video, but that's before you consider that this treatment could be applied to any static images that already exists. So maybe ten years from now you'll go on Spotify, check a certain box in the options menu, and suddenly every album cover is moving around to some degree or another, sometimes specifically adjusted by-hand by the artist, sometimes automatically determined by an algorithm. I'm pretty sure that's what's going on with those "1950's Super Panavision" videos that are getting popular on youtube right now, except I'm suggesting that on some platforms there could be no still images, everything could be alive with motion So when we are each able to cultivate such algorithms with intentionality and artistry, that I think that will be a fundamentally different artform which is equal to photography in its impact. And I think we're going to have something similar with AI music eventually, where it is more procedurally-generated in the moment rather than static file retrieval. And I think there are more possible totally unique artforms that we will see emerge from this tech, especially as they start to be combined, like how photography progressed to video with sound. That's hard to say, but I think generally it will push people to be more original. I think the idea of just having talent for talent's sake will lose value, and what will matter more is what new ideas people bring to their work. It will not be as interesting to just make something pretty. As for what it does better, well that depends on whether you compare it to the abilities of individual artists or all artists. It can obviously do a lot of things better than any one individual artist. Comparing against all traditional artists, I think we can generally say it is much better at brainstorming rough drafts of developing concepts quickly, we can get fully-formed pieces rather than just relying on sketches. That's just one example. Ok, yes, that is a fair point. So I asked if these "social media spats" were relevant, and on second thought, yes they are. That is unfortunate, but actually I think an argument in favor of DS and other underground music/art scenes beginning to use these tools openly. That is too much power to be utilized solely by toxic weirdos. We need people who venerate beauty, originality, love, etc, to be at least as visibly and prominently working with AI I think. I mean, I don't think those values are under threat, just that we can try to make the future look beautiful instead of bleak, why not? But about whether the early primitive AI art will have the same temporal strangeness as early digital art, early 3d, early sampler music, etc. I think for sure yes! For example, I always think back to this Pouff – Grocery Trip video for when the visual AI stuff first started making splashes with DeepDream, back when it always seemed to generate amorphous dog faces. And for the current era, yes I personally will probably forget about all the memes, anime waifus, and Greg Rutkowskis that the tech bros churn out, but perhaps those things are important and meaningful to those guys and they might fondly look back on in the future. I think that's good, let more people experience the joy of creativity, even when it's not something I will be able to appreciate. I know I will be happy to look back on my own Midjourney generations, even if nobody else ever has any interest in them (perhaps we could start an AI art thread...?). There's nothing to stop people making art in the holodeck dreamscape, it could be done traditionally, by prompting, or even just mental intentionality like a lucid dream. There can be as much or as little human input into the experience as you want, or you could just simulate living in prehistoric times or whatever. I don't think it will be societal suicide, but instead will coincide with the death of society as we know it, like post-singularity. But I think the in-between stages to eventually reach this vision will be very interesting. It's not a guarantee that we'll get there though, which is why I think it's important to try and make this vision clear with art, or some other positive vision of our potential long-term future, so that we have something to strive for and aren't just afraid of what's to come. in a niche genre where people do everything out of their pockets asking them to pay sometimes 100 +++ (!!!!!!!!!!) euros/dollars for a cover art that could be WORSE than the dreaded AI... What would suck more about that than spending the money is telling the artist that you're deciding not to use their cover and instead commissioning someone else, or feeling obligated to stick with cover art that you don't like just because you don't want to hurt their feelings. Brutal. Is it just me or is point 5 a pretty good point? That it breaks past-fantasy immersion because it's futuristic? I suppose it's subjective, but I don't see why AI should necessarily break immersion any more than using synths. It depends on how you use it I think, like a crusty old mellotron sample fits perfectly with the dungeon vibe, but a gloopy analog Moog-type synth patch I often find to be immersion-breaking.
I actually kind of agree with this. I am all for remix, sampling etc. I think a computer "remixing" a human's artwork is worse to me than a human doing it. This is entirely based on how it feels to me, but hey we're talking about art so feelings are relevant. It feels slightly frightening in a science fiction post-human terror kind of way that a robot can remix someone's artwork with minimal input, sell it to someone else (because a lot of AI art is paid for with subscriptions etc.), and the artist not even know about it or get compensated. It feels fundamentally different to me because the process is automated by code, and I for one think it feels slightly scary that human considerations such as moral, ethical or legal considerations are either non-existent or being outsourced to machine-learning. This is not an argument, it is not me saying that all AI is bad or that no one should use it. But it is the feeling I get having looked into how it works more recently. This is why I'm going through the process of updating all my covers to be public domain, because it just makes me slightly uncomfortable. The copyright issue, and the ability to rip off specific artists' styles, I think those are real problems for sure. I think there are solutions though, like showing your work. I think this can apply to both traditional artists and AI artists, like posting videos of the creation process for traditional art, and for AI artists by providing the prompt/model/seed/etc. Also I think it would be reasonable for legislation to happen now that prevents training on data without permission (not saying I want that to happen necessarily, just that it is definitely worth considering). But also I think an expectation that AI art be noncommercial is a good way to keep it separate and maintain a perspective of the relative values of these things.
|
|
|
Post by zorkdrrtolk on May 24, 2024 18:13:21 GMT -5
"There is absolutely no human involvement behind it." You say that like it's a bad thing. Have you seen humans lately? Definitely not a very savory species. Allied Mastercomputer, is that you?
|
|
|
Post by moreteeth on May 25, 2024 6:16:54 GMT -5
I think that drawing a connection between photography and artificial intelligence is a false equivalency for more reasons than one. The camera was a revolutionary device because it was an entirely new art form, and that conversely changed how classical art was made; movements which highlighted the craft of painting itself—impressionism, for example—were created because the previous functions of painting as an accurate replication of reality had been supplanted. In essence, photography was an entirely different thing that served a similar function. However, artificial intelligence as its currently used is entirely derivative. I was not trying to draw an equivalency between AI art and photography, just using it as an example of a disruptive new technology. Maybe the drum machine would be a better example? That was mostly derivative and bemoaned by people who thought it would cause harm to them, but we would not condemn everyone who used a drum machine. There are good and bad ways to use it like any tool. I do think AI art and photography have a lot of parallels though, but I actually think AI art is more revolutionary. The key phrase you said, "as it's currently used..." I admit I cannot think of a clearcut example where it has been used in a way that is totally novel, but I can imagine ways it might be eventually. For example, what if, when the models get efficient enough and the compute cheap enough, we could have "images" that are constantly being generated (like 24+ fps)? So let's say you had a landscape of a tree on a cliff near the sea during a storm, the wind could be blowing the grass and leaves, the tree swaying, the branches rustling, the clouds moving, the waves crashing against the shore, etc. Or let's say you had a photorealistic portrait of someone on a busy street in New York, the person could be smiling for the camera, sometimes laughing, looking around, shifting their position, scratching their nose, and all the while behind them there are cars and people constantly passing through the frame, never repeating. This might seem like no big deal, like just watching a video, but that's before you consider that this treatment could be applied to any static images that already exists. So maybe ten years from now you'll go on Spotify, check a certain box in the options menu, and suddenly every album cover is moving around to some degree or another, sometimes specifically adjusted by-hand by the artist, sometimes automatically determined by an algorithm. I'm pretty sure that's what's going on with those "1950's Super Panavision" videos that are getting popular on youtube right now, except I'm suggesting that on some platforms there could be no still images, everything could be alive with motion So when we are each able to cultivate such algorithms with intentionality and artistry, that I think that will be a fundamentally different artform which is equal to photography in its impact. And I think we're going to have something similar with AI music eventually, where it is more procedurally-generated in the moment rather than static file retrieval. And I think there are more possible totally unique artforms that we will see emerge from this tech, especially as they start to be combined, like how photography progressed to video with sound. That's hard to say, but I think generally it will push people to be more original. I think the idea of just having talent for talent's sake will lose value, and what will matter more is what new ideas people bring to their work. It will not be as interesting to just make something pretty. As for what it does better, well that depends on whether you compare it to the abilities of individual artists or all artists. It can obviously do a lot of things better than any one individual artist. Comparing against all traditional artists, I think we can generally say it is much better at brainstorming rough drafts of developing concepts quickly, we can get fully-formed pieces rather than just relying on sketches. That's just one example. Ok, yes, that is a fair point. So I asked if these "social media spats" were relevant, and on second thought, yes they are. That is unfortunate, but actually I think an argument in favor of DS and other underground music/art scenes beginning to use these tools openly. That is too much power to be utilized solely by toxic weirdos. We need people who venerate beauty, originality, love, etc, to be at least as visibly and prominently working with AI I think. I mean, I don't think those values are under threat, just that we can try to make the future look beautiful instead of bleak, why not? But about whether the early primitive AI art will have the same temporal strangeness as early digital art, early 3d, early sampler music, etc. I think for sure yes! For example, I always think back to this Pouff – Grocery Trip video for when the visual AI stuff first started making splashes with DeepDream, back when it always seemed to generate amorphous dog faces. And for the current era, yes I personally will probably forget about all the memes, anime waifus, and Greg Rutkowskis that the tech bros churn out, but perhaps those things are important and meaningful to those guys and they might fondly look back on in the future. I think that's good, let more people experience the joy of creativity, even when it's not something I will be able to appreciate. I know I will be happy to look back on my own Midjourney generations, even if nobody else ever has any interest in them (perhaps we could start an AI art thread...?). There's nothing to stop people making art in the holodeck dreamscape, it could be done traditionally, by prompting, or even just mental intentionality like a lucid dream. There can be as much or as little human input into the experience as you want, or you could just simulate living in prehistoric times or whatever. I don't think it will be societal suicide, but instead will coincide with the death of society as we know it, like post-singularity. But I think the in-between stages to eventually reach this vision will be very interesting. It's not a guarantee that we'll get there though, which is why I think it's important to try and make this vision clear with art, or some other positive vision of our potential long-term future, so that we have something to strive for and aren't just afraid of what's to come. I think your hypotheses regarding the potential of artificial intelligence are interesting, but you must admit that it's somewhat speculative, no? All these things sound great in these hypothetical futures and within these incredible contexts, but as of right now it isn't looking too hot. Maybe I'm not enough of a visionary; I don't know. In regards to your thoughts on the holodeck dreamscape, I think that you consider artificial intelligence as more of a "key" to future forms of expression more than the premise behind its conception. You think less of a "artificial intelligence holodeck" and more of just a "normal holodeck" where artificial intelligence might serve as its foundation. This also circles back to my comment on artificial intelligence and its function: why make me do laundry and an artificial intelligence do art when it can be the other way around? Sure, it'd be great if—in the far future—artificial intelligence made a post-singularity simulacrum for us to live and do art in, but does that future necessarily require all the other things which artificial intelligence is being used for? As always, the technology itself is fine, but the ethics of its use in the hands of the laymen is of chief importance. Although the technology of the early internet may've facilitated that primitive style of web design, the aesthetic extremes in the examples I referenced weren't in any way inevitable, while the aesthetic progression of artificial art will be linear and subject to the whim of the technology's progression. Sure, there have been some minor changes and innovations regarding the technology behind painting or music—for example—but the same instruments can express entirely different qualities over (baroque, classical, romantic, minimalist, etc.) While in some ways the earliest Deep Dream generations possessed a brief novelty, artificial art is ostensibly assessed on how well it can replicate normal art; in other words, the more something looks like artificial intelligence, the less it's valued. I do like your drum machine comparison, because they were essentially seen as derivative representations of actual acoustics drums, but then were purposed in industrial, techno, etc. However, not only is Dalle already nearly capable of representing actual art, but the awkward stages its been up until now have been anything but charming. In a way, the frontier has already closed. Not only that, but I can think of litany of albums which would be made worse by exclusion of the drum machine but I can't think of a single work of art which would be worse if it wasn't a product of artificial intelligence.
|
|
|
Post by magicwillow on May 25, 2024 9:05:52 GMT -5
The copyright issue, and the ability to rip off specific artists' styles, I think those are real problems for sure. I think there are solutions though, like showing your work. I think this can apply to both traditional artists and AI artists, like posting videos of the creation process for traditional art, and for AI artists by providing the prompt/model/seed/etc. Also I think it would be reasonable for legislation to happen now that prevents training on data without permission (not saying I want that to happen necessarily, just that it is definitely worth considering). But also I think an expectation that AI art be noncommercial is a good way to keep it separate and maintain a perspective of the relative values of these things. I totally agree with this. I think your last point about AI art being relegated to noncommercial is a great idea.
|
|
|
Post by andrewwerdna on May 25, 2024 14:23:10 GMT -5
I think your hypotheses regarding the potential of artificial intelligence are interesting, but you must admit that it's somewhat speculative, no? All these things sound great in these hypothetical futures and within these incredible contexts, but as of right now it isn't looking too hot. Maybe I'm not enough of a visionary; I don't know. Yeah, most of what I've been saying is speculative for sure, but I just want people to consider that maybe this all might be going somewhere beautiful and good, and we're more likely to get there if we can approach it with an adventurous optimism. I am not understanding the distinction between “AI holodeck” and “normal holodeck.” I think you're still operating under the assumption that with AI art, the AI is doing pretty much everything. Yeah, if you have an android in your home that does the laundry that's great, and if it likes to make oil paintings in its free time I also think that's great (and I personally would be interested to see what it's making on its own volition). But if you sit down next to your android and ask it to “paint a scene of elves parading through a mystical temple on the moon, combining the styles of Hieronymus Bosch, Caravaggio, and John Bauer,” and then work with it for hours every night for a week, adjusting the prompt, tweaking the android's parameters, iterating through many subsequent variations, and doing many other spontaneous modifications, perhaps even adding a few brushstrokes yourself, then I think the human is “doing the art” as much as the android. It is just a very new and unfamiliar way of “doing art.” That future I'm envisioning doesn't require all the the things AI is currently being used for, but I think it certainly requires the ability to make art, music, video, chatbots, etc. I think those things are even required right now and can't wait until later, because right now I think it is through those things that we can try to recognize what the AI is and understand it to some extent. If we treat it as just an out-of-sight out-of-mind drudgery servant then I think we are going to be totally unprepared and blindsided when AGI emerges. But also it's just astonishing and magical, and I don't think we should deny ourselves these aesthetic experiences just for some futile attempt to cling to a status quo that hardly anyone is satisfied by. I think people will always be making art by hand, and yes it might become more rare as AI gains greater adoption, but that is not necessarily a bad thing. Traditional handmade art may have even more value than it does now because of that relative rarity, even if it might seem humble in comparison. And of course consider how much more valuable live music will be once we're inundated with AI music. I think those primitive web design examples, while they may not have been an inevitable result of technology's progression, they can still be described as “subject to the whim” of it. I do not think the aesthetic progression of AI art has been linear or will be going forward. It is just hard to get a handle on right now because these things are so complicated and happening so fast, but that's why right now is such an opportunity, most people are clueless about what to do with this stuff so they don't even bother. If we can manage to have some idea of what we're doing and just start working then that alone will put us very far ahead, and we will maybe be able to set a lot of the standards that are built upon in the future. After all, look at the competition, we could wipe the floor with what those unimaginative tech bros are doing, we just need little bit of space to share and discuss our work. I totally think there is charm to the awkward stages! For example here's some stuff I made in Artbreeder a few years ago that I still think looks pretty cool: www.artbreeder.com/vhacyrkan?sort=new It will need time, but I'm sure eventually that primitive AI art aesthetic will be appreciated and explored further for its uniqueness. Can you think of many AI artworks that you've been able to appreciate? It is so early that we have not had time to make the works, to appreciate the works, and for many people to even allow themselves to appreciate the works. Also many works can only be made through collaboration with AI, impossible to conceive of making it traditionally. Hell, even those primitive Artbreeder images I linked above, what would it even mean to “exclude the AI” from those? Now that I think about it, it always has been a completely new art form, and photography is the best metaphor we have. It is not derivative like a drum machine, not at all, only if people intentionally use it in a derivative way.
|
|
|
Post by andrewwerdna on May 27, 2024 22:26:10 GMT -5
Anyways... sorry for ranting so much about this. I get hyperfixated on things sometimes, but I think I got out pretty much everything I was needing to say, so I will shut up about AI now.
|
|