There's quite a few people on here using them, but from what I've seen it was mostly for portraits for stuff like visual novels.
The main limiting factor keeping it from being more widespread, imo, is animation (because, and as harsh as this is going to sound, nobody really cares that much about the ethic side of things right now except artists or people in the online art sphere). Most AI I've seen is just now starting to get a good grasp of making static images, but the stuff in terms of animating a small video is uncanny and a bit choppy. With most games, that probably isn't going to improve much unless someone feeds a model some good ol' post-the Dover Boys cartoons in order to teach it how to stretch frame (It'll eventually become very adept at frame-by-frame posing for more rigid designs, I have no doubt, but for things like weapon trails or non-"realistic" artstyles it'll still need to know how to do convey motion in a single image). And even then it's going to take a relative while for it to learn how to pull that off in a convincing way.