Ok but what is an actually good solution to the whole AI art debacle?

eldritch_lich@lemmy.world to Ask Lemmy@lemmy.world – 129 points –

I heard about C2PA and I don't believe for a second that it's not going to be used for surveillance and all that other fun stuff. What's worse is that they're apparently trying to make it legally required. It also really annoys me when I see headlines along the lines of "Is AI the end of creativity?!1!" or "AI will help artists, not hurt them!1!!" or something to that effect. So, it got me thinking and I tried to come up with some answers that actually benefit artists and their audience rather that just you know who.

Unfortunately my train of thought keeps barreling out of control to things like, "AI should do the boring stuff, not the fun stuff" and "if people didn't risk starvation in the first place..." So I thought I'd find out what other people think (search engines have become borderline useless haven't they).

So what do you think would be the best way to satisfy everyone?

61

I think if capitalism wasn’t involved in how AI evolves, we would be in a much better place. The fact that the first question about any tech is “how can we make money with it?” Already starts down a different path.

Seems like since we can’t solve that fundamental issue, best next bet is to learn to welcome AI into our lives in ways that enrich it. Use it to augment your work - alternatively, maybe start learning and specializing in things that are (for the time being) out of reach for AI. Human services that require another human or hand made high quality items where the purchaser is specifically interested in the hand made aspect.

Can’t say I have a perfect solution, other than to stay curious and adaptable to change.

This is a big part of the issue.

The truth is that most artists would still make their art if they didn't make money off of it provided that they are ensured basic necessities and enough time.

The whole idea of AI art would be much easier to brush away as a non-issue for many artists if it didn't immediately pose the question of "how is this going to affect my ability to live?"

I wholeheartedly believe this too. There's something so amazing about the feeling of creating things with your own hands and seeing what the rest of the world says about it. But the moment you rely on this to literally not starve, any unfairly advantaged competition becomes that much more dangerous.

Yes, indeed. And, in fact, it would probably make them happier if the bullshit commercial art that's bereft of meaning - a subset of commercial art, to be clear - was no longer something they had to concern themselves with in order to pay the bills.

If they - and the rest of us - could be granted the dignity of a good and comfortable life while perusing projects that were personally fulfilling, then there'd be little issue. But instead, their work is being used not only to line the pockets of investors and a handful of tech enthusiasts with the right connections (generative machine learning models aren't that hard for anyone with some programming skills and basic linear algebra to develop, the magic is in the money), but to directly undercut their paid work. And that's just not going to fly.

I went to art school, and I distinctly remember people talking about art being one of the few things that was safe from AI. "They'll have computers driving cars and doing office management but they can't do anything creative so it's going to be a good time to be an artist" and so on.

I guess you could make the argument that they're not really being 'creative' right now, but if the output is good enough for large amounts of the general public then it's still just as damaging for artists I think.

We have an art director at the corp I work for who is basically automated out of a job now. She is near retired and on good terms with the CEO so they will probably let her wrap up and not replace her.

Of course she still hasn't figured out how to use the new label maker, update the website, or use AI art generation so that has fallen to the rest of us.

Counter argument to pose, would be that a skilled artist with AI is now a faster producing artist than without - presumably (at least at the current tech), this combo pair up is best of both worlds. Artist can create art but still retains creative freedom and the talent of guiding AI prompts in specific directions a project may call for that a non-artist with an AI would struggle with.

I'm an artist - I tattoo, do freelance illustration and produce handmade pottery. My husband is also a tattoo artist. My entire income is made through art.

I have stopped attempting to draw coloring books - AI "prompt artists" have taken over and are pumping out grayscale coloring books at extremely low prices. Not a high income producer for me in the first place, but the entire field is falling apart.

Tattooing is a different story - I use AI to produce references regularly. Not full drawings, just references I can use to create my own drawings. Pottery remains unchanged.

The obvious difference is the type of art. The further it moves from a drawing, the better the outcome when AI is involved from my POV.

To be honest though - how many of you actually have real artwork in your house? Not prints - actual handmade art. Art has been struggling for a long time now - it has little value to the average consumer. Mass production has made it a throwaway product. Most ceramics are made by machines now - vases and "paintings" and dishes are all isles in a home goods store, stamped out and inked by a machine. Most professional artists are employed by companies, not off selling their art. I don't really need to spell out what will happen when the company gets a hold of a free program to replace their artists.

There isn't a good outcome for artists here - consumers want cheap art. Companies want cheap artists. Artists want living wages and for a lot of us that means not making a living off of art already, because the wealthy class that has luxury money to spend on handmade and original art is shrinking as we speak.

At least - it is here in America.

it has little value to the average consumer.

The average consumer can't afford it. I can't afford to pay a living wage for 40-80 hours worth of work to put on my wall.

I'm happy to say I have real art in my home, but now that I'm thinking about it... if I didn't live with an artist, there would be a lot less of it.

Universal basic income.

EDIT: I feel I need to clarify this stance. I'm a game developer, and I'm also sometimes an artist. I've been in discussions with my artist friends and they're of the belief that AI is the enemy. I don't believe that, but I also don't blame them; artists' careers are being displaced by AI, so of course they are justified in directing their anger at the immediate threat to their livelihoods. As for the ethical/legal quandary about AI art models being trained on publicly-available copyrighted images, well, I think that's more of a grey area that I should not delve into right now.

AI art is inevitable, and there is no feasible way to slow it down, let alone stop it; even if the developed Western world decide unanimously to ban AI art, black markets for AI art models would still exist, and it would give Eastern countries (okay, mainly just China) the opportunity to really corner the market and further develop that technology for (nef)various purposes.

Every industry succumbs to automation sooner or later. I mean, it makes perfect economic sense to adopt automation to replace the existing workforce; robots don't need to get paid. However, normally, when automation is introduced into an industry, the pace at which that technology is developed and implemented is slow enough for displaced workers to be able to learn new trades and pick up new careers before they, well, go broke.

The issue is that in this era of information, this AI technology is developing far too fast for that to happen, and couple that with the ever-increasing greed in today's capitalism that ensures workers (including artists) are working long hours and are being paid barely enough to survive; they don't have the time, energy, or funds to learn new trades or adapt to other creative endeavours that haven't been taken over by AI yet.

We need a return to an era where art is not a method of livelihood, but rather an admired craft, simply for the purposes of art; that also comes with the extra bonus of lowering the worth of AI art. For all that to be possible, artists first need a way to stop worrying about having to make money to be able to afford food and housing.

That's where UBI comes in. Again, capitalism, the richest 1% hoarding 50% of the world's wealth, yada yada, just tax those rich fucks and give that money to everyone; boom, UBI. Once artists (and workers in general) no longer have that threat of starvation, 1. they can take all the time they need/want to learn new crafts and careers, 2. they can force corporations to offer workers better treatment and pay, especially when they can afford to quit toxic work environments.

Focus on physical art. Sculpture, screen prints, paintings on canvas, mixed media, etc.

Trying to stop progress has never worked. Sure, there are regulations that should be put in place, but putting the genie back in the bottle isn't going to happen.

The same debate happened when photography was invented. Photography led to some of the most iconic art movements of the 20th century and the deliberate departure from more realistic paintings, because nobody needed to hire a painter or sketch artist just to remember what someone looked like anymore.

My expectation is that AI will continue to be used to generate what has already been identified as particular styles. And these styles will go out of fashion with overuse, much like the 50's was oversaturated with mass produced designs. Which led to pop art.

Tl;dr It's all just part of the creative development of humans and our adaptation to automation. If we want to address why artists starve, AI in art is only a symptom of a much larger issue around human worth.

The AI has particular styles because the current AIs are trained in a particular way. As we go further, there is no reason to assume that AI would not be able to copy/mimic any stile. While photography is distinguishable from painting, the AI painting can be indistinguishable. That’s, I think, is fundamental difference.

A painting can be distinguished from a photo from its physical dimensions and attributes like texture. If I take a photo of a painting, almost nobody is going to file a court case against me for putting it in a book or website or tv. We're collectively ok with that level of reproduction.

An AI work can currently be determined by looking for 'errors' that aren't likely to be drawn in by artists or found in reality. Inconsistent shadows, exact repetition of organic pattern details, extra limb bits, doesn't construct language fluently, etc. We're not necessarily ok with that level of reproduction though because it is slightly less obvious at first glance.

Just like photos are a lossy format for paintings, AI is a lossy reproduction tool. While it will definitely be harder to distinguish with time, we're sort of still left with the core problem. What fidelity reproduction is necessary for something to no longer be original? And what does the originality actually matter when people don't need to rely on producing it to eat?

Hell, people even bought NFT apes with cut and pasted Mr. Potato Head parts for thousands. People have photocopied photos of the Sistine Chapel in books for close to free. Maybe it's just cool that the imagery is resonating with people regardless of the medium.

Also, we should feed people without the expectation of financial return, thanks for coming to my tired tedx talk.

If AI gets to the point where it can truly replicate anything, even so it won't be the end of human creativity, so I expect humans will still be making new things and new forms of art.

If AI can replicate everything, it will be able to replicate that as well. The only way to get around is to have something like “sertifiably human”. Like today, most chess competitions are for humans only.

I strongly agree with the comparison to photography. People like Man Ray and David Hockney opened doors to new art impossible before photography. And it did not eliminate drawing, painting and similar arts. It did force some change upon them, though. It seems that we have discovered something like a new art form. That should be exciting. It will certainly be uncomfortable

There is no easy solution, hence companies trying to push extreme tactics like C2PA. I'm not even convinced at all about it, because people will definitely figure out how to fake image metadata. It'll also take ages for it to become a standard, and it can't be a worldwide standard when you know some people will reject it.

Realistically the best option is accepting that AI art is a thing and at the very least making sure it stays open source. It would be terrible if only large companies would have access to this tech. Meanwhile artists can make something like a timelapse video to prove their art is genuine. Some people are also guessing with the rise of "artificial" art, art IRL will be way more valuable like theaters, statues, paintings etc.

I don't like to be a pessimist, but as a musician and writer, and seeing how the value of my work has steadily decreased years before AI became mainstream, I don't see how "real" art will become more valuable. Maybe on an individual/personal level, but not in the industry as a whole. Especially once an untrained person can't tell the difference between AI and "handmade" art.

I think the slew of flops at the box office is a sign that people are rejecting this. Putting aside what AI could do, right now everything that's generated feels vapid in a sense, and I don't think that's entirely because it came from a machine. The creators were just that uninspired.

I'm personally happy that the film industry is struggling while works like Spiderverse and Helluva Boss and going against the status quo.

I mean, there's a reason that Marvel/DC comics are nowhere near as popular as they used to be while manga gets several aisles at the bookstore.

It still hurts to see people around me lose motivation because of AI though.

AI art is like the knock off movie studios. It uses someone else's ideas and rehashed it in the most generic way possible because that is what it is designed to do.

At a certain point the AI could probably make a convincing time lapse. Just thinking out loud, it really is a tricky problem

Considering the amount of processing power needed to make a decent AI model, I'm pretty sure it's already solely controlled by large companies. Plus, if it becomes legally required then people can't exactly reject it.

In my personal opinion, I don't think AI art is inherently bad and I'd put it on the same level as that particular style of soulless corporate art. I'm confident that people who actually care about the quality of whatever it is they're making will commission real artists. And the existence of AI art wouldn't take away the enjoyment of creating something with your own hands. But I'm not a professional artist so I think my opinion is irrelevant anyway. If actual artists have a problem with it, then it needs to be addressed.

While I mostly agree with you in that there's no way most people would be on board with C2PA, it's an entirely different matter if it becomes legally required. I don't know how likely it is but it doesn't seem impossible.

(Also the impersonation argument feels contrived to me. Just get your info from the source 4Head)

Well, it just bothers me that I know many people who still think art and other creative pursuits should be relegated to hobby status and I should get a "real" job. And the fact that AI is doing things that humans are supposedly meant to do for fun just doesn't sit right with me.

Considering the amount of processing power needed to make a decent AI model, I’m pretty sure it’s already solely controlled by large companies.

Well, yeah, but not exactly. The best image generation model is Stable Diffusion which is open source. There's a huge open source community vastly improving image generation and creating amazing features for it that outpace companies. Stable Diffusion XL is almost done and also set to be open source. There's also a big push for language models to be open source, but we're not there yet.

the fact that AI is doing things that humans are supposedly meant to do for fun just doesn’t sit right with me.

I totally agree but sadly it's where technology has lead us. It turns out making image/text/software generating AIs are so much easier than robots that automate the boring stuff. Physical robots aren't there yet. I don't think computer scientists intended to destroy art, but more just "this seems like a logical next step that AI can do".

The big problem just lies in money. Millions of people will lose their jobs quicker than they think over AI advances, and it'll be a slow transition until we can create an economic system that can sustain them. The "just get a real job" crowd are in for a rude awakening when they realize there will be no "real job" 10 years later.

But ah well, I'd encourage people to just enjoy the rollercoaster ride and see how it goes rather than shouting at computers.

As another user pointed out, as long as capitalism has control, there will be no good solution.

I think about it this way: does knowing art is AI generated take away from the experience?

  • If the answer is yes, then it simply will never take over. As long as we have some sort of law that requires art to be tagged as AI if it's AI generated, then I think that would be enough. No need to tag original (human) art with anything, no need for that kind of surveillance, just tag AI art or make companies legally required to divulge if it is so.

  • If the answer is no, then I think this is just the natural progression of things. There will always be artists, and there will always be people that want their art to be made by a person. But if most people really don't care if there's a person behind the brush, then it doesn't matter if it's AI or not.

I don't think anyone has a right to what they do. If you're an artist, that's all good, but if your art isn't appreciated, if people prefer AI over your art, then why should we block AI? Just so you can keep making money off your drawings? There's other things you could be doing... Once again capitalism makes it so that hobbies can't just be hobbies, "if you're not making money you're failing", so this isn't a very satisfying perspective, but the reality is that you don't need to be just an artist, you can have a job and draw for fun, post things online, etc.

Ideally, we tell capitalism to fuck off. We already produce enough to feed, clothe, house, and heal all of the world, we don't do it only because the oligarchs choose not to, as it's not profitable. If we set up a system that actually makes use of the value generated by labour, instead of letting the 1% hoard it, then AI would not be threatening any job security, and it wouldn't be stifling creativity or anything, it would just be a tool.

There is the issue of copyright though, since original works are used to train AI. That whole debacle is a can of worms that I will not open.

If the answer is yes, then it simply will never take over. As long as we have some sort of law that requires art to be tagged as AI if it’s AI generated, then I think that would be enough. No need to tag original (human) art with anything, no need for that kind of surveillance, just tag AI art or make companies legally required to divulge if it is so.

I find it hard to imagine that such a law would ever be implemented, even harder to imagine it would be enforced well, and even if both of those happen, that companies wouldn't find ways around it like just oversaturating everything with "may contain AI generated imagery" so that the tag becomes entirely meaningless.

adding a mandatory label doesn't really feel like a good solution to me.

There is the issue of copyright though, since original works are used to train AI. That whole debacle is a can of worms that I will not open.

You had some great points, but then glossed over the biggest issue we have, as artists, with AI. People’s entire portfolios were scraped to train these things without their permission.

First thing I learned in art school is that theft and forgery are the most lucrative careers in the arts. (In hindsight, maybe a red flag there.) If an AI developer came into an artist’s studio, grabbed all their paintings and ran out, you can clearly see that’s theft. Just because the art theft is automated and largely unseen, doesn’t mean it isn’t still theft. People don’t like having their life’s work stolen from them.

Generative AI could have been a great artistic tool, but it’s been tainted for a lot of artists by the blatant, egregious theft at its very core.

Yeah sorry, I just don't know enough about the technology and copyright law; I didn't want to make uneducated statements. I'll read the article you linked though, and thank you for your input!


::: spoiler Just one thing...

If an AI developer came into an artist’s studio, grabbed all their paintings and ran out, you can clearly see that’s theft.

With this, the artist is losing possession of their art. AI doesn't take art away, it creates more art based on original works, right? It's more akin to plagiarism than physical theft, though of course plagiarism is very much theft so I think I get your point.

AI tools need to be integrated into all workforces alongside appropriate training.

One of the hurdles for AI that I don't see being overcome soon is the fact that all of AI is regurgitation. It cannot generate an original statement. Thus the focus should be how AI can enhance your workload. How do you train your AI models to increase your productivity at an individual level?

It’s a very difficult topic, and I don’t see any satisfying real-world solutions. Two big issues:

  1. Obvious solutions are impossible. Generative AI are impossible to "undo". Much of the basic tech, and many simpler models, are spread far and wide. Research, likewise, is spread out both globally and on varying levels from large Megacorps down to small groups of researchers. Even severe attempts at restricting it would, at most, punish the small guys.

I don’t want a world, where corporations like Adobe or Microsoft hold sole control over legal "ethically trained" generative AI. However, that is where insistence on copyright for training sets, or insistence on censored "safe" LLMs would lead us.

  1. Many of the ethical and practical concerns are on sliding scales. They are also on the edge of these scales. When does machine assistance become unethical? When does imitating the specific style of an artist become wrong? Where does inspiration end and intellectual rights infringement begin? At what point does reducing racial and other biases from LLMs switch over to turning them into biased propaganda machines?

There are dozens of questions like these, and I have found no satisfying answers to any of them. Yet the answers to some of them are required in order to produce reasonable solutions.

Hopefully liberalisation - let AI do the tasks that it can do.

Generating 3d models and sprites will liberate thousands of indie film-makers and game developers - just like cheaper computers and cameras have made it more accessible too.

It's a tool like any other - we should embrace it for productivity and lowering the barrier to entry.

Also for copyright and IP - we really need reforms to abolish software patents globally, greatly reduce the extent of copyright (to ~10 years) and likewise for patents. This will encourage the creation of more novel works to get copyright, and also allow for a massive growth in the creative use of existing popular works (and for training AI models).

I think copyright should go back to it's original terms in the USA - 17 years by default, with an optional 17 year extension.

I had 85+ images of my art used to train AI. I think the best solution is for the current AI image training sets to be cleared and rebuilt on copyright free and opt in only content. Similar to stock photography where artists can decide for themselves if they want minimal compensation to contribute their art to the training set. This would be necessary because once the systems have been trained on an image it's in the memory. So the only way to respect the rights of the artists after the fact is to wipe the computers and start image generation all over again but ethically. I have linked on my Mastodon https://www.youtube.com/live/uoCJun7gkbA?feature=share a senate hearing on the issue in which the lawyer from Universal Music perfectly pointed out "it'd be hard to opt out if you don't know what has been opted in" . Additionally, this isn't just an artist issue the training set includes photos from medical records, schools, and personal photos. Basically if you've ever posted a photo on the internet there's a chance it's in the training set. "have I been trained" is a website where you can see what is included and opt out (though as mentioned earlier that's not a good solution) I spoke to a prominent IP lawyer in Chicago (before the class action lawsuits were public) and he pointed out that they didn't have the right to reproduce my artwork into their training set. Their actions have been likened to a smoothie shop. They have the storefront and the blenders but they stole all the ingredients. After it's blended you may not ALWAYS be able to recognize the strawberries BUT we know they didn't pay for the fruit. It was stolen for their profit. Why should I be forced to provide the core product of my business to develop the core product of another (for-profit) business?? The senate hearing linked above includes many other important and valid points. Myself and many other artists I know aren't against AI. I love tech and think it's really fun and can be helpful, it just needs to be done ethically. I have a lot more I could add to this, hahahaha

Its hard cause even that analogy falls apart, lots of people believe piracy isnt stealing, because the original still exists in its entirety - and with web scraping legal in the US, by putting that image out there you're indirectly allowing other artists to take ispiration from it - so what morally draws the line between biological and artificial inspiration? Its a moral delemia I havent yet heard a good solution to personally

I mean there is an actor and writer strike for this reason and more. AI is not art. While it's true that art is subjective its also supposed to invoke emotion and get a person to think. Not that AI art can't do that, but it's not the same.

I think AI art is great for first concepts, practice, really anything but the end result.

Hm, can you elaborate further? I don't think you've supported in your point in that you say that AI art can achieve the same subjective outcome of invoking emotion and getting a person to think, but you concluded that it's not the same.

I feel like there's a finer point you want to make but haven't gotten across yet.

some art is meant to get you to think
some art is meant to fill a gap on a wall

i don’t think a poster from ikea is meant to get you to think. a live laugh love slogan isn’t even creative (any more; maybe the first one was?)

there’s lots of things that visuals are meant to do, and get you to think is just 1 of them

I agree, but alot of people don't see it that way. As an example, Sakimichan is a very popular artist, they have been active for like 20 years now, maybe more. As a result they had alot of art on the web. Their style was shnthed a lot, without their permission. If you generate art, you're very likely to get something that looks like their art as a result.

They cannot keep up with the instant result of AI art synching their stuff. They used to just do tasteful pinups. Now they are doing full on porn in order to compete with the AI blight on their career. And it's very obvious to me that they do not enjoy drawing that kind of stuff.

Sakimichan is not the only one seeing the fallout of AI art, a lot of artists are having trouble. And it is getting harder and harder to tell what is made by AI. Then you have photoshop's thing that can enhance your art with AI...whatever thst means. Art communities have joked for years that PS had a "make art better" button when asked how people draw so well. Now it actually has one.

I'll agree that AI isn't art, in the same way a paintbrush isn't art or photoshop isn't art. It's a tool that can be used to create art. I think a fascinating application of it is when Corridor Digital worked it into a creative pipeline to make their Anime Rock Paper Scissors video.

That being said, the shareholders don't see AI that way. They see it as a way to copy the unique art style and aesthetic of another artist without paying that artist for the years they spent getting good at making their art.

True artists create their own pigments from scratch and harvests their own minks for brushes. Anyone who uses pre-made paints is a fraud!

AI is another tool and while it won't be artistic on its own, art can made using AI just like any other tool. Just like Clipart, spiralgraphs, and paint can be used to make quality art or just random crap.

The main problem is the copyright infringement since it is basically a more technical version of tracing someone else's work.

In any case, governments should work on strengthening social programs and safety nets. Consider basic income and ideas to mitigate job loss, perhaps adding extra tax to businesses that replace humans with AI to fund those programs. You do not want thousands of people angry and hungry on the street tearing you down.

AI needs data to train up on. It can't create art without first consuming existing art and spitting out parts of the originals. There's a reasonable claim to be made that AI synthesis of prior art is itself original enough to count as having intrinsic worth, but if the only way to get it is stealing other people's work to train up your model, the whole value proposition of AI art is probably net negative, entirely at the expense of artists whose work was used to feed the model.

Yes, there's the argument that automation of new things is inevitable, but we do have choices about whether the automated violation of copyright to feed the model is tolerable or not. Sure, it's a cool sexy technology and who doesn't love getting on the bandwagon of the future and all, but the ethics of modern AI development are trash and despite promises that automated AI labor will save the owner class money by doing for free what the plebes demand to be paid for, it's really as much a ponzi scheme as all those crypto currencies that don't have intrinsic value unless enough suckers can be convinced to feed the scheme.

And yet, it's a powerful technology that has potential to be a legitimate boon to humanity. I'd like to see it used to do things (like picking crops that are hard to automate with dumb machines, or cleaning trash off of beaches or out of the ocean, or refactoring boilerplate code to not use deprecated packages or to review boilerplate contract text for errors) that aren't just ways for owners to cut labor out of the economy and pocket the differences.

Perhaps, if we are going to allow AI to be a great big machine that steals inputs (like art, or writing, or code) from others and uses them to do for-profit work for their owners, the proceeds attributable to AI ought to be taxed at a 90%+ rate and used to fund a Universal Basic Income as payment for the original work that went into the AI model.

AI art is art, period. Just like with any method of creation, there will be good and bad AI art, and as with any method of making art, there is human input and intention behind it. Internet is chock full of same-looking fan drawings of popular characters—everyone can pick up a pencil, do a 15·minute sketch of Joker; or grab a camera, shoot a landscape, and upload it on Deviantart. Same for boring, uninspiring, mass-produced commercial art.

Fundamentally generative neural networks are no different from "oldschool" procedural generation tools like Mandelbulb3D or Terragen—with both of which I have tinkered a lot in the past. With AI you use a verbal prompt to generate; with "oldschool" generative processes you use a numerical input or different math formulas.

As for AI somehow "stealing" art, well, every artist who studies the works of other artists to learn how to make good art, is "stealing", then. At the end of the day, a human brain is literally a neural network that can be trained using various inputs. No input; no output other than random noise. From my own past a decade ago tinkering with digital art—one of my renders with Mandelbulb that was well recieved on DA (ended up in some curated collection, even) was based on someone else's input formula that I tweaked heavily and used different render parameters—and I'm sure someone else took my version of that formula and made their own version.

That's the nature of art, nothing is created in vacuum; nothing is original. Every artist "steals". Those who claim different, who believe art should never draw from other art, are either delusional or pretentious elitists. Or lawyers.

I think ensuring growing funding for arts should be less directly connected to the work they create. One side of this is abolishing interleectual property rights. The other side would be a proper UBI + probably art/culture endowments and alternative revenue structures.

How to get there without a bit of civil war I'm not so sure.

Maybe training images should only come from consenting artists and photographers

So what do you think would be the best way to satisfy everyone?

Your barrel of thoughts is not wrong: current AI is just a tool, if people were already satisfied in their life they wouldn't be bitching about it but unfortunately everyone is struggling and competing to pay rent or make a living or worst to stack money to compensate to their dissatisfactions.