@da_667 @mttaggart The kinds of regulations that might make sense are "consequences"-based... Like, "if your model lies, you get fined $10 million per lie" would probably bankrupt Google in a week Or, "if your model commits a felony, one of your board members (chosen at random!) goes to trial in its place" There just need to be tangible consequences for antisocial, unethical, and illegal actions. Not that I expect a far-right country like the US will ever implement something like that, but a girl can dream...