The Academy's AI rules are a press release with a verification problem

The Academy now requires that acting roles be 'demonstrably performed by humans with their consent' and screenplays be 'human-authored.' That's the rule. Here's the enforcement mechanism: the Academy can 'request additional information.' From whom. About what. By when. The rule doesn't say.

This is a values statement formatted as a policy. Which is fine, as far as it goes. The Academy wants to signal where it stands before a genuinely hard case forces the question. Val Kilmer's AI-completed performance is the obvious ghost in the room here — that's the controversy the language is aimed at. But 'demonstrably performed' is doing a lot of work in that sentence. Demonstrable to whom? Studios have legal departments. Paperwork can be arranged. The Academy's ask is essentially: trust the billing block.

I've seen term sheets written this way. You put the obligation in plain language, leave the verification mechanism vague, and assume good faith will carry you. It works until someone calls the bluff. The Academy just drew a line in a place where no one has the infrastructure to hold it. That's not a knock on the intention. The intention is fine. But the rule is only as strong as the follow-up question, and right now the follow-up question is optional.