Skip Navigation
153 comments

  • I've always been confused about this train of thought, because it seems to justify the opposite of what it's trying to say.

    I mean, if the argument is people will use whatever garbage they have on hand to make art... presumably that includes generative AI? Look, I lived through four decades of people making art out of ASCII. My bar for acceptance for this stuff is really low. You give people a thing that makes pictures in any way and you'll get a) pictures of dicks and b) pictures of other things.

    I don't think GenAI will kill human art for the same reasons I don't think AI art is even in competition with human art. I may be moved or impressed by a generated image, but it'll be for different reasons and in different scales than I'm... eh... moved and impressed by hot dragon rock lady here. Just like I can be impressed by the artistry in a photo but not for the same reasons I'm impressed by an oil painting. Different media, different forms of expression, different skill sets.

    • Nothing will kill art itself, GenAI will simply be incorporated as another tool

      Killing the ability to make money from art AND the bs that corporations are pulling in regards to AI, profit and making line go up is what people are mad about, but that anger is constantly misplaced leading to lines of thought like this lol

    • I think the argument is that an AI "artist" is incapable of creating art. Their "tool" does the work for them. Whereas other artists use digital tools but as just that - tools. The art comes from the artist.

    • In the absence of needing to use skills to make a living, I have no problem with AI art. In a hypothetical anarchist mutual aid society, people could make art with whatever methods they prefer. Some might create AI models to make art because they're interested in that sort of thing. Others will make art in the traditional ways, also because they're interested in that sort of thing. There doesn't have to be tension between the two, and their basic needs are all there.

      When people have to use their skills to make a living, though, then there's a problem. So many of the places that were paying artists are now whipping something out with an AI model. That leaves artists without a way to cover their basic needs at all.

      • I don't know how much that logic tracks, at least long term. And I don't know that I'm going to be more inclined to be on the side of human labor over automation now when I wasn't for garments, car manufacturing and other commodities. The John Henry of visual arts I am not.

        I do have a couple of seemingly opposing but not contradictory points to add to that, though. One is that historically anti-automation, anti-industrialization movements have a pretty bad track record at succeeding. The other is that I think you're giving "AI art" way too much credit. Small and medium-sized commissions may get impacted (I am on record saying that AI is the new "cousin who knows Photoshop" and I stand by it). For anything an actual professional needs to book and hire based on quality? Nah.

        There may still be an impact on that high end, because I expect that generated elements will become a tool in an artist's toolset more than anything else. That may speed work up and require fewer people, but not "leave artists without a way to cover basic needs" necessarily. Just like photography, just like CG, just like Photoshop and so on. There was doom and gloom around all of those as well, and hyperbolic claims from tech peddlers, too. Go look up some of the claims of early photography entrepeneurs about what the technology would eventually be able to do, some are hilarious.

        I also expect sooner or later people will get good at spotting telltale machine-generation quirks and put additional value in organic, human-looking creative products. People are already misidentifying human art as AI art, artists will likely lean into that. Think vinyl into CDs back into vinyl or the premium on less processed foods more than... I don't know, cars that don't have rattling doors or whatever.

        That's a guess or a forecast, though. We'll see where it goes.

      • When people have to use their skills to make a living, though, then there’s a problem.

        Progress leaves many professions behind. It's lamentable, but a price worth paying.

    • This pretty well encapsulates my feelings, except for the issue of training the models. AI is cool tech, but the fact remains that people are making money off of scraped content. Not to mention the environmental aspect.

      Honestly I find it difficult to reconcile.

      In a perfect world, we would have open source models trained on public domain and properly licensed content.

      I don’t think AI is going to replace artists any time soon. On the personal side, people create for the joy of it, whatever that means to them. On the professional side, people have a hard enough time communicating what they want to an actual person, much less a computer.

      As someone that likely has moderate aphantasia, I really struggle with describing what I want. Being able to tell an image gen to make so many variations of X, and then commission a friend to take inspiration from Y and Z to make something original is really freeing for both sides, imo.

      I’ve never gotten exactly what I’m looking for, but it almost always gives me something to point to, without doing a bunch of test drafts. I suppose that’s technically taking work away from the artist, but so does having an ‘undo’ button in procreate.

      Idk, it’s a more complex issue than many make it out to be. I’m still further on the fuck ai side than not, just due to its current implementations.

      End rant.

      • I mean Adobe firefly addresses the properly licensed dataset issue and afaik it's all viewable (though I'd much prefer something anyone could use offline locally). Environmental impact will always be an issue unless we see some evidence of mitigation either from direct green energy use or at least creating additional green energy generation from any organization doing the base model training.

    • The thing is, an AI 'artist' isn't making art. They are generating images with no real meaning or effort put into them.

      • That depends on what they're doing. If they're entering a prompt and rolling with what they get out of it, then sure.

        If they're inputting a prompt and refining it with solely AI tools then meeeh, that starts to fade a little. I'd ask why someone is spending hours going back and forth with an AI instead of doing some of it manually, but it's hard to tell one way or the other from the final output.

        If they're inputting a prompt, refining it with AI tools and heavily editing what comes out in image editing software that's approaching some strange digital mixed media weirdness I don't think we have particularly good intuitions for.

        If they're inputting a prompt and using the output as some building block like a texture on a 3D model or for a content aware fill in photo editing or for a brush or a stamp I genuinely have no mental model for what impact that has in my assessment of the "meaning" or "effort" going into a piece, if I'm being perfectly honest.

        Reductionism isn't serving us particularly well on this one. Makes the pushback feel poorly informed and excessively dogmatic.

  • The future is approaching. When society will collapse a new Furry-Stone age will begin...

153 comments