The biggest AI labs now have a place to register with New York, and a deadline. OpenAI and Anthropic must sign up by 1 January 2027, when the state’s RAISE Act begins to apply to them.
Only companies earning over $500 million a year are covered. Their duties: post a safety plan online, alert the state within three days when something serious goes wrong, and publish model reports. The penalty is up to $1 million for a first offence and $3 million for repeats.
The state has created a new agency, DIGIT, to enforce all this. TechCrunch covered the portal’s opening on Tuesday and was told by a state spokesperson that action against labs that skip registration is possible, though no specifics were given.
At the same event Governor Kathy Hochul opened a research centre at Binghamton University, in the SUNY system, that will test advanced models independently and advise DIGIT on the labs’ safety plans. She first announced it in January with $55 million behind it. Her pitch was aimed at tech leaders who call their AI plans inevitable and treat harm along the way as acceptable.
The rules are part of a patchwork. Compare California, where SB 53 gives labs a looser 15 days to flag critical incidents. Illinois will add outside audits when its SB 315 starts next January. New York also already enforces a law telling AI companion apps to remind users they are talking to software and to look for signs of self-harm, and a chatbot law for minors, signed in October, applies from 1 June 2027.
Pakistani users will not see a different ChatGPT or Claude. The difference is on the labs’ side, because more of their safety paperwork becomes public.
