TechCrunch reported that New York opened a sign-up portal on Tuesday for its RAISE Act. Frontier developers, the firms building the most advanced AI models, must register by January 1, 2027. The law covers companies with over $500 million in yearly revenue, such as OpenAI and Anthropic. They must post a safety framework, a written plan for handling model risks, and report critical safety incidents within 72 hours. They must also publish transparency reports on their models.
A new state office, DIGIT, will run the law. Penalties top out at $1 million for a first violation and $3 million for each one after. The state can bring enforcement actions against firms that fail to register, but a spokesperson declined to give specifics. TechCrunch does not say what counts as a critical incident or what the documents must contain. DIGIT and a new Binghamton University center will review the frameworks.
The likely benefit for buyers is that safety claims move into public documents they can read and question. Because the required contents are unknown, buyers can ask vendors how they define a critical incident and what their framework covers. Only companies above the revenue line are covered.