What the RAISE Act provides
New York has opened the registration portal required by the RAISE Act, the state law on AI safety and transparency that will take effect at the beginning of 2027. According to TechCrunch, large frontier developers—that is, companies with annual revenue above $500 million—must register by January 1, 2027; the source cites Anthropic and OpenAI among them. The law also requires them to publish their safety frameworks on their websites, report critical incidents within 72 hours, and produce public transparency reports on their models.
Enforcement will be entrusted to the new Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT), created to oversee large frontier labs. The penalty framework indicated by the source provides for up to $1 million for a single violation and up to $3 million for subsequent violations. The state can also initiate enforcement actions against those who do not register on the portal, but a spokesperson did not provide details on the matter.
The new evaluation center and the regulatory context
In parallel, Governor Kathy Hochul announced the launch of the New York Center for AI Responsibility and Research, based at Binghamton University, part of the State University of New York system. The center will independently evaluate the safety, security, and operation of advanced AI systems and will collaborate with DIGIT in reviewing companies' safety frameworks. Hochul had announced the structure in January, with $55 million in donations and capital.
At Tuesday's briefing, the governor touted the law's approach, arguing that it is a regulatory model that puts communities above corporations and people above technology; she also criticized the tech executives' narrative that their vision of AI is inevitable. The RAISE Act is not New York's only intervention: the state already enforces a law requiring companion AI operators to periodically remind users that they are talking to a bot and to make reasonable efforts to detect suicidal ideation and self-harm. In October, Hochul signed another law on chatbot safety for minors, which prohibits behaviors such as deceptively human-like interactions, emotionally manipulative or flattering responses, and encouragement of conduct such as suicide or eating disorders; it will take effect on June 1, 2027.
Comparison with other states and open questions
The source places the RAISE Act alongside other state laws. In California, SB 53 took effect in January and requires frontier labs to adopt and publish safety frameworks, introduce whistleblower protections, and report critical incidents within 15 days. In Illinois, SB 315 will be operational from the beginning of 2027 with similar requirements, plus independent safety compliance audits.
Some aspects that the material does not specify remain to be clarified: what concrete sanctions will be applied to those who do not register, how reports within 72 hours will be verified, and by what criteria the Binghamton center will evaluate the frameworks. The framework, therefore, is defined in its main outlines, but its practical implementation will depend on DIGIT's procedures and the announced collaboration with the new research center.
