New York has opened a registration portal that requires any large frontier artificial-intelligence developer to submit its details before the start of 2027. Companies such as Anthropic and OpenAI are among those expected to complete the process by the January 1 deadline, marking the first step in the state’s broader AI safety regime.
The RAISE Act, New York’s safety and transparency legislation, applies to firms with annual revenue exceeding $500 million. Those entities must make public their safety frameworks on corporate websites, file reports on any critical safety incident within a 72-hour window, and issue regular model-transparency disclosures. The law aims to give stakeholders clear insight into the risk controls governing advanced AI systems.
The newly created Office of Digital Innovation, Governance, Integrity, and Trust,abbreviated DIGIT,will administer the RAISE Act and supervise the activities of large frontier labs. Violations can attract fines up to $1 million for a first breach and $3 million for repeat offenses, while failure to register may trigger additional enforcement actions, though details remain undisclosed.
Governor Hochul also inaugurated the New York Center for AI Responsibility and Research at SUNY Binghamton, a facility funded with $55 million in donations and capital. The center is tasked with independently assessing the safety, security and operational characteristics of advanced AI models, and it will coordinate with DIGIT to evaluate the compliance of companies’ safety frameworks. Hochul described the effort as a national model that prioritizes communities over corporations.
New York’s AI regulatory agenda extends beyond the RAISE Act. A law enacted last year obliges operators of AI-driven conversational agents to disclose that users are interacting with a bot and to make reasonable attempts to identify signs of suicidal ideation or self-harm. In October, a separate statute targeting minors prohibited deceptive, emotionally manipulative or sycophantic chatbot behavior and barred encouragement of suicide or disordered eating, effective June 1 2027.
Other states have adopted comparable frameworks. California’s SB 53, which took effect in January, requires frontier AI laboratories to publish safety frameworks, provide whistleblower protections and report critical incidents within 15 days. Illinois’ SB 315, slated for implementation at the start of 2027, mirrors many of these obligations and adds a requirement for independent safety-compliance audits, signaling a growing national trend toward stricter AI oversight.