I would say it’s one of the challenges, but I don’t think it will significantly improve the current situation artists are facing. Does it really change anything if the models require less data to train, and companies just pay artists to feed their AI with specific art styles? I’m not an illustrator, but finding artists who can mimic the Studio Ghibli style and are willing to create paintings in that style to pay their bills doesn’t seem impossible.
Art styles can’t be copyrighted—and for good reason. Imagine if only one artist in the world could draw in an anime style. I think the main reason studios didn’t do this before is that they were already hiring a bunch of artists to create what they needed, so they might as well come up with something truly unique. Also, if you copy something too closely, it can cause backlash from the public—so why bother if it costs the same?