I've played a bunch of great RPG Maker and Ren'Py games that used creative commons art and music. The draw of those games was experiencing the story that the creator wanted to tell. What's the difference if the same creator used AI generated art instead? Seems to me it could help creators to express themselves better, not worse.
But in your case the art would be in everything else that isn’t the sprites. By everything else I mean everything the author would have care about. It wouldn’t mean that the sprites couldn’t be aesthetically appealing but only that this part of the game wouldn’t be what’s artistic about it.
I think IA can be incorporated into art as long as it is made consciously, but IA generated content in itself couldn’t be the art piece.
When a game maker collaborates with human artists who contribute their individuality and self-expression to the final product, it’s amazing. When it supports those artists, either via commission fee or as a credit and word of mouth, it’s amazing. When game maker finds people who genuinely want to hear the story they want to tell, it’s amazing.
All of those amazing things are fading away when a game maker could instead be shelling out some bucks (not a lot for now, while they operate at a loss) to Microsoft or Nvidia, until a Frankenstein monster trained on the unlicensed works of those very artists spews out something passable.
Then, tomorrow there comes out a product that generates a complete game on demand (probably trained on scraped, unlicensed source code and assets) and puts that game maker in the shoes of the aforementioned artists. Good luck finding people who want to hear the story.
I wish I could believe that this will somehow result in creativity flourishing, but I am yet to see a convincing take on how this would happen. All I can foresee is creativity being stunted.