I see this thread got double posted and I posted in the less popular one so I'll put this here too:

I’m about as in favor of AI art as can be but I don’t think apathy is really the right reaction. It will impact the careers of a lot of artists, but it’s not going to be artists literally becoming obsolete overnight like everyone pretends it is.

For most of it that we’re used to seeing, we’re seeing art from people who were not going to pay an artist for what they made in the first place. People running pictures of Yellow Parenti through an anime model were never going to pay an artist to do that for them, without AI image generation they’d just… not do it. The artists would be no better off. This is literally the same shit record and film companies try to argue with piracy, that one download = one lost sale. That idea is bullshit in that context, and it is bullshit in this context.

The other main area we might see AI art more widely adopted in would be any type of art or photography that is repetitive but covers a broad range of subjects, like stock imagery. I can’t speak for people who work for stock image companies but I cannot imagine that being anything but soul crushing as-is, posing for and making shitloads of photographs of the most inane and specific things. AI still has a lot of catching up to do before it can consistently replace that, especially with human subjects, so there’s still time.

I think most people don’t think of the limits because most people do not have any experience with it beyond seeing whatever people put out with a smartphone app, then seeing moderately clearer images later on, and then they assume that it’s going to be replacing all artists by the end of next year. I have a fair amount of experience by this point with making my own AI art and I can tell you that regular artists will absolutely continue to be a thing for quite a while even in the worst case, if nothing else because AI is so terrible at making specific new subjects from scratch (at least without a LOT of manual effort) and I cannot imagine it getting much better any time soon. I’ve been trying to make an original character, with some fairly complex detailing, 100% from AI art with no prior references existing. In order to do that to any degree of usefulness, I have to make probably 50-100 completely flawless pictures of my character in varying poses, zoom levels, art styles, and backgrounds – I can do that with the AI, but it takes a lot of time. It took me an entire WEEK to get one picture completed, and I’ve only recently got remotely useful results straight out of the generator. The only reasons I am willing or able to do this are that I have the hardware to do it, I have a decent understanding of computer science, I have some very basic art skills, and I have a strong aversion to social interaction – plus I just like the novelty of the idea. Not everyone has all of that and I am the only person I know of that is attempting this. Most people will just either pay an artist for this instead (a vastly cheaper option, I assure you) or just not do it at all. For the foreseeable future, getting a specific thing out of an AI will take far more effort than 90-99% of the population is willing to expend compared to paying an artist to do it for them.