New York opened a portal on Nov. 10 for large frontier AI developers to register under the Responsible AI Safety and Education Act. The state’s new Office of Digital Innovation, Governance, Integrity, and Trust runs it within the Department of Financial Services. Large developers have until Jan. 1, 2027, to register, the date the amended law takes effect. The governor’s announcement and the chapter amendment

The state uses two thresholds. A frontier model is a foundation model trained with more than 10²⁶ operations, counting specified subsequent training and material modifications. The additional obligations for large developers apply when a frontier developer and its affiliates together had more than $500 million in annual gross revenue the preceding calendar year. Those obligations include keeping a disclosure statement current and publishing a safety framework. The department’s guidance for developers

A smaller frontier developer can still have an incident to report. All covered frontier developers must publish model transparency reports and notify the state of qualifying critical safety incidents within 72 hours of determining that an incident happened or learning facts sufficient to reasonably believe it did. A developer must also disclose an incident posing an imminent risk of death or serious physical injury to an appropriate authority within 24 hours. The department’s reporting guidance

Put the registration in the compliance calendar and the incident deadline in the response plan. Before Jan. 1, confirm who updates the disclosure if model ownership changes and who can notify the state when an incident starts on a weekend.