Not directly related to Vtubers themselves, but to the meta-discussion around AI art that keeps popping up;
So apparently a team at the University of Chicongo have developed a method for making your online art "poisonous" to machine learning, making it junk data so it gets rejected by the AI learning algorithm. Honestly this would be a good solution for everyone if this got mass adoption (and if it works as well as it claims to do), artists don't have to worry about their works being used and AI devs don't step on anyone's toes.
Personally i would find it fascinating for one reason; if this got mainstream would we see a drastic reduction in quality in AI art if their samples decreased, how necessary is "unauthorized" sampling of works for machine learning really? This and other solutions might be something to keep an eye out for in the future, this particular programme i found due to Vtubers and their fans retweeting the launch announcement so it's already had some penetration into our sphere.