I have some issues with these strawmen arguments.
"But what if you sell it? What if you don't credit them at all? What if you had a machine make the collages? What if you made 1000 collages per minute? What if one of those collages became wildly popular, and sold for more than their 10 salaries combined?"
Comparing this to music for a moment, we can ask the same thing of sampling. If I like an artist that samples work, am I obliged to go explore the original work? Well, the answer imo falls into how much is changed from the original work. Paying respects to the original is good practice, sure, and is something that every fan of sampling should do, but, to use a real-world example, liking the glitch-trance of Everywhere at the End of Time doesn't mean that I need to start developing an affinity for 40s contemporary ballroom jazz.
If the piece is different enough for the original to warrant existing, and yet isn't presented as an entirely original work (read: it's honest about being derived from other works and created through the heavy usage of some program), I'm not going to be bitter.
And as far as the "throw shit at a wall until it sticks" argument goes, yeah, that pisses me off, but that's more of a problem in that attention is warranted as monetary gain in today's information economy moreso than a fault of AI. To use another example, if all those stupid toy/Elsagate channels on Youtube become fully automated, I'm not going to suddenly be pissed that they're not made by humans, I'm going to be pissed that Youtube let their system get so lazy and predictable that it can be that easily gamed. "Don't hate the player, hate the game."
"AAA studios can use AI art too, yeah? I agree that saving indies money is nice, but it's only achieved through cutting indie artists out of the gamedev space."
Yeah. Indies and AAA have the same tools to produce similar products. That's what "level the playing field" means.
And it doesn't really cut indies out of gamedev space. In fact, there are VNs I've seen on here that use AI art to supplement character portraits. As much as I dislike them personally, I can't fault the method since its just someone using a tool to supplement for their lack of specialized skill elsewhere. It's the exact same thing as someone using a pre-made game engine instead of coding their own.
That strawman argument isn't even true when we look at outliers that only appear in the AAA scene. Very few indie games, if any, have access to something as robust as the Fox Engine, or character designs as recognizable as Nomura's (though the latter can be replicated). And that's to say nothing of the marketing budgets that AAA has access towards! AI doesn't take these types of resources away from AAA. So really, it's not changing the status quo because AAA still has those types of technological advantages and more.
I'd like to end by highlighting that AI derives from patterns that exist, and tries to derive from what it's learned. A human can still compete in that regard so long as it's art has something to stand out. I have zero doubts that(, say,) Temmie Chang's art could be replicated by an AI (given three years), but the pieces would still ultimately be recognized as "Temmie Chang," which ends up advertising her work so long as she would be active enough to remain in public consciousness otherwise. The secret to survive in a post-AI marketplace would be the same as our scene now with several thousands of humans each trying their own style - being distinguished enough in a work that's happenstance popular would proliferate the artist, human or not, above the rest of the white noise of competition.