New York starts policing the biggest AI companies on January 1

New York is getting ready to enforce a major AI safety law. On Tuesday, the state opened a registration portal. Large AI developers such as OpenAI and Anthropic must sign up by January 1, 2027. After that, they will have to be open about how they handle safety, and they can be fined if they are not.

What the law requires

The law is called the RAISE Act. It covers large frontier developers, meaning companies that build the most advanced AI systems and earn more than $500 million a year. Under the law, these companies must:

• Publish their safety frameworks on their websites.
• Report critical safety incidents within 72 hours.
• Issue public transparency reports on their AI models.

Who enforces it

A new state office called DIGIT, short for Office of Digital Innovation, Governance, Integrity, and Trust, will carry out the law. Penalties can reach $1 million for one violation and $3 million for later ones. The state can also take action against companies that fail to register. A spokesperson declined to say what that would look like.

Governor Kathy Hochul also launched the New York Center for AI Responsibility and Research at Binghamton University. It will independently test the safety, security and operation of advanced AI systems and work with DIGIT to review company safety frameworks. She first announced it in January with $55 million in donations and capital.

Other rules in the mix

New York already enforces a law for AI companion apps. Operators must regularly remind users they are talking to a bot and make reasonable efforts to detect suicidal thoughts and self-harm. A separate chatbot law for minors, signed in October, takes effect June 1, 2027. It limits deceptive human-like behavior and emotionally manipulative responses. California and Illinois have similar frontier AI laws. California's took effect in January and allows 15 days to report incidents. Illinois's starts in 2027 and adds independent audits.

What it means for you

If you use or build on popular AI tools, the companies behind them will have to say more about safety and report serious problems. What is not yet clear is how strictly New York will enforce the registration rule, or which companies will end up covered beyond the examples named.