Given how new AI is to the world, and our criminal law system and society really, hasn't adapted to AI

Until it has, and maybe even after it does, we should continue to regulate or at least monitor AI use, to a extent, even if it is supposedly not harmful.