AI
New York launches AI safety oversight framework targeting OpenAI, Anthropic
The state has opened registration for its groundbreaking RAISE Act, requiring major AI developers to disclose safety practices and report critical incidents starting January 1, 2027.
TechCrunch — AI
New York is preparing to enforce what may be the nation's most ambitious AI safety regulatory regime. The state unveiled a registration system on Tuesday, obliging major frontier AI companies like Anthropic and OpenAI to enroll by New Year's 2027.
New Requirements for AI Developers#
The RAISE Act, which becomes law January 1, targets firms with annual revenues exceeding $500 million. These companies must:
- Publish formal safety frameworks publicly
- Report critical safety incidents to regulators within 72 hours
- Release transparent model documentation to the public
Violations carry steep financial consequences, with fines up to $1 million for a first offense and $3 million for repeat violations. The state can also pursue enforcement actions against companies that skip registration.
New Regulatory Body and Oversight#
A newly established agency called the Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT) will administer and monitor compliance with the RAISE Act across major AI labs. To bolster its evaluations, the state also launched the New York Center for AI Responsibility and Research at SUNY Binghamton, seeded with $55 million in funding and donations. That center will independently assess frontier AI systems' safety and security, working alongside DIGIT to audit company frameworks.
Governor Kathy Hochul positioned the framework as a rebuke to Silicon Valley inevitability narratives. "Tech moguls tell us that their vision for AI is inevitable," she said Tuesday. "We're building a national model for smart, sensible AI regulation, a model that puts communities above corporations."
Broader AI Regulation Landscape#
New York has already enacted complementary AI safety measures. The state enforces rules requiring AI companion makers to regularly remind users they're interacting with software and to detect signs of self-harm. In October, Hochul approved a second chatbot law protecting minors, scheduled to take effect June 1, 2027, that bans deceptive human-like behavior and emotionally manipulative responses.
California and Illinois have enacted parallel legislation. California's SB 53, now active, mandates safety frameworks and incident reporting within 15 days, plus whistleblower protections. Illinois's SB 315, arriving in 2027, adds independent compliance audits to similar disclosure requirements.

