New York opened a registration portal on Tuesday for the RAISE Act, its AI safety and transparency law, according to TechCrunch. Developers of frontier models with annual revenue above US$500 million, including OpenAI and Anthropic, will have to register by January 1, 2027.
Under the law, these companies must publish safety frameworks on their websites and meet transparency requirements. The report calls the rule one of the most significant of its kind in the country. With enforcement set to begin early the following year, state oversight moves from announcement to practice.
For companies that run large models or use them in their products, the practical effect is a requirement to document safety practices publicly and in a way that can be verified. Those who rely on these models also gain access to information about which protocols their providers have said they follow.
The Novidades da Semana view is that the requirement to publish frameworks establishes a minimum standard of accountability that was missing. The revenue threshold, however, leaves out smaller firms that also build advanced models, and that gap should be watched as the law is applied.
