I expect the next major copyright law legal battle will be around the question of whether content generated by some machine learning approach that was trained on copyrighted data is a derivative work or not.