I think the actual concern people have with AI generated art is that the bar for entry has been lowered (expected) but we're ignoring the attention to detail or verifying/confirming the output. Too much trust is being put into AI generated output and we rush to publish stuff that doesn't convey the original idea. With it being commercially driven art, the incentives aren't there to properly review when you are emphasizing quantity over quality.
Take the swan and army boat in Los Angeles lake example. Before fully AI generated art existed, a human would have looked and reviewed that and ask the artist, "why the hell is there a swan there?" There could be a dialogue to defend or change that.
Even without AI, this problem still would have existed if your goal is producing as much distinct art at scale with less humans in the loop. Things are bound to slip through but it only winds up being smaller issues that might not impact expression of idea. Adding fully generated AI art has amplified this problem to a point where we don't have time to review at all.
The politics of the people taking that stance are all over the map, and the arguments are about personal economics and immediate human impact.
You may think it's all nonsense, but I don't know how you get to "left-wing" or "culture war" from that.
No.
The problem is the economic impact on artists.
I’m just saying the training data issue is a dumb hill to die on. Because it will get sorted out eventually and the “best case” is some loose change thrown at a current generation of artists.
It is an ineffective argument if your goal is to help artists.