Award bodies are writing the AI governance that legislatures won't

The Golden Globes just published eligibility rules for the 84th ceremony that ban performances substantially generated by AI, require full disclosure of any generative AI used anywhere in a submission, and prohibit unauthorized use of a performer's digital likeness, voice, or biometric data. The Val Kilmer trailer situation is in the background here. So is the 2023 strike. The rules are specific. They have a compliance mechanism. They define what counts as human authorship.

That last part matters. Most responsible scaling policies currently on the books at major AI labs don't do that. They describe intentions. They gesture at principles. They were written with a press release in mind and a lawyer in the room. The Golden Globes policy, by contrast, tells you what disqualifies a submission, who has to authorize what, and what the threshold is for a human remaining the credited centerpiece. That is a functional governance document.

This is where we are. A film awards body owned by Dick Clark Productions has clearer, more enforceable AI rules than the U.S. federal government. The Academy guidance aligns. The industry pressure is real. And the actual frontier labs — the ones building the systems that will make all of this much harder — are still operating under scaling policies drafted for a different audience entirely.