I don't see why ML tools need to understand this from the law's perspective. The publisher & content producer already has responsibility to verify. There are al

I don't see why ML tools need to understand this from the law's perspective. The publisher & content producer already has responsibility to verify. There are already tools that allow an author or artist to sue if the work is too similar, or to request a publisher to do a takedown. If generative ML cannot sufficiently distinguish itself from a blatant copying machine, then we already have a system in place to deal with that problem. But there seems to be the real possibility that ML tools can create art that's sufficiently different, and the degree to which they can generate sufficiently new art is the degree to which they warrant a new policy framework. IMO the question of an artist's copyright ownership is the wrong framing here with regards to fear of ML. If self driving technology gets better, taxi and truck drivers do not have any informal or formal ownership over the art of driving, but they have the same fear.