I expect relying on people to ground and proofread AI to be as immensely successful as policies inspired by “Guns don’t kill people, people kill people.” People are the problem. People will continue to be the problem. Policies need to account for actual human behavior and not some ideal.
John ScalziWe're years into "'AI' will absolutely positively make up shit, never use it for anything even vaguely important" being common knowledge and yet this keeps happening, at a certain point you just have to accept some people just plain like being found out and humiliated www.ft.com/content/a61c...