New York Rolls Out Strict AI Safety Rules for Big Labs
Frontier AI firms must register, share safety plans, and report incidents fast under the RAISE Act.
TL;DR
- NY opened a portal requiring large AI labs to register by 2027 under the RAISE Act.
- Companies over $500M revenue must publish safety frameworks and report incidents in 72 hours.
- New DIGIT office oversees rules with fines up to $3M; a SUNY center will run independent tests.
New York opened a registration portal this week for one of the stricter AI oversight plans in the US. Big frontier labs like OpenAI and Anthropic now have until January 1, 2027 to sign up.
The RAISE Act targets companies pulling in more than $500 million a year. They must post safety frameworks on their sites, file public transparency reports, and alert the state about serious incidents inside 72 hours.
A new Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT) will watch over compliance. Miss the rules and fines can hit $1 million for the first slip and $3 million after that.
Governor Kathy Hochul also launched the New York Center for AI Responsibility and Research at SUNY Binghamton. More than 80 state scientists will run independent tests on advanced models using compute from the Empire AI partnership.
“The stakes are too high to simply take AI companies at their word,”
“The stakes are too high to simply take AI companies at their word,” Hochul said.
The center gets backing from $55 million in donations and capital. It will check safety claims, study risks, and feed findings to officials. The law still only covers what companies choose to disclose publicly.
California and Illinois have rolled out similar transparency requirements, though their timelines and exact rules differ slightly from New York’s.
The Tea
States are tired of waiting on Washington and are writing their own rulebooks. Labs that treat safety as optional now face real paperwork and potential fines if they skip the line.