Jargon Busted: The RAISE Act
New York's RAISE Act makes the biggest AI labs show their homework.
TL;DR: The RAISE Act forces the largest AI developers to publish safety plans and report incidents fast, and it takes effect January 1, 2027.
For anyone staring at the latest AI governance headlines and thinking RAISE Act, what now?, here’s your five-second decryption.
New York’s RAISE Act makes the biggest AI labs show their homework. If you’re a “frontier model” developer (training systems on more than 10²⁶ FLOPs and pulling in over $500 million a year), you now have to publish safety protocols and report serious incidents to the state, some within 24 hours. Think of it as California’s frontier AI law with New York’s own teeth.
So what’s this got to do with your feed full of AI headlines? Its proof states aren’t waiting on Washington.
Why it matters:
Only a handful of companies are big enough to qualify, but those giants now answer to a New York regulator, not just their own PR team.
It lands mid federal preemption fight. The White House wants one national rulebook. New York just showed that states will legislate anyway.
Penalties bite: up to $1 million for a first violation, $3 million after that.
Watch this space closer: the federal preemption fight is heating up, and New York just became Exhibit A in the states’-rights argument. 👀