New York is now at the forefront of AI governance after Governor Kathy Hochul signed the Responsible AI Safety and Education (RAISE) Act into law, a bold move that imposes safety and reporting requirements on large artificial intelligence developers — even as the federal government moves to limit state regulation of AI.
This landmark bill shows how states are taking charge of AI safety in the absence of detailed federal rules, setting a potential blueprint for other jurisdictions seeking to balance innovation with public protection.
A New Model of AI Oversight at the State Level
The RAISE Act — signed on December 20, 2025 — requires AI companies that generate over $500 million in revenue to establish comprehensive safety plans for their most powerful models and publicly disclose how they prevent their systems from causing critical harm.
Beginning January 1, 2027, companies must also report serious safety breaches within 72 hours or risk significant fines. Enforcement will be managed by a new office within the New York State Department of Financial Services, which will issue regulations and publish annual state reports on AI safety compliance.
These provisions mark one of the most detailed AI regulatory moves by a U.S. state to date, expanding on ideas first championed by California in its landmark Transparency in Frontier Artificial Intelligence Act (SB-53), which similarly requires transparency from AI developers on risk assessments.
What the RAISE Act Covers — And Why It Matters
The RAISE Act is targeted at so-called “frontier models” — advanced AI systems developed at high computational and financial cost that could pose existential or catastrophic risks if misused.
Under the law, developers of these models must go beyond voluntary internal safety measures and submit structured plans detailing how they:
- Assess and mitigate critical risks such as the autonomous creation of harmful biological or chemical agents.
- Monitor and prevent their systems from engaging in criminal conduct without meaningful human oversight.
If a serious problem occurs — like a model’s unexpected behavior that could endanger lives or property — companies must notify the New York Attorney General and state emergency services within a 72-hour window, a significantly tighter timeline than similar laws in other states.
Federal vs. State: The Clash Over AI Authority
New York’s action comes amid growing tension between state governments and federal policy on AI regulation. In 2025, President Donald Trump signed an executive order seeking to curb states’ ability to independently regulate AI, arguing that a patchwork of rules could stifle innovation and hinder U.S. competitiveness in the global AI landscape.
Rather than step back, New York officials publicly rejected the notion that states must stand down. Assembly member Alex Bores, the RAISE Act’s sponsor, argued that waiting for federal action risks leaving citizens unprotected. He noted that when federal leadership stalls, state-level intervention becomes not just helpful — but necessary.
This conflict echoes similar debates in other states, like California, which has also pursued its own AI safety legislation and criticized federal attempts to pre-empt state oversight as counterproductive.
Reactions from Industry and Advocacy Groups
Tech industry players and lobbyists raised concerns throughout the RAISE Act’s negotiation, warning that overly prescriptive requirements could hamper innovation or create regulatory uncertainty. Many encouraged aligning the law with California’s SB-53 standards to promote consistency across state borders.
Meanwhile, parent groups and public safety advocates have pushed back against those efforts, arguing that strong safeguards are essential — especially given AI’s rapid integration into everyday life and the potential for misuse in areas ranging from fraud to biological threats.
What This Means for AI Development in the U.S.
New York’s RAISE Act may have far-reaching implications beyond its borders. By setting a precedent for state oversight of AI that includes enforceable safety regimes and strict reporting timelines, it could spur other states to follow suit or push Congress toward more comprehensive federal legislation.
In the meantime, developers of powerful AI models — including companies working on generative AI, autonomous systems, or large language models — will need to prepare for an emerging regulatory environment where state and federal policies intersect and sometimes conflict.
In the absence of unified federal rules, New York’s approach demonstrates how elected officials are attempting to protect public safety while still fostering innovation — a delicate balance that will continue to shape AI policy debates nationwide.
A New Chapter in AI Governance
As AI systems continue to evolve and integrate deeper into society, governments must decide who sets the rules and how compliance is ensured. The RAISE Act marks a significant milestone in that journey, challenging the notion that AI can be left entirely to industry self-governance or deferred exclusively to federal regulation.
New York’s bold step reflects growing recognition that AI’s potential benefits come with risks that merit structured oversight, transparency, and accountability — even if it means states charting their own course. For more on AI regulation trends and the broader policy landscape, see Regulation of Artificial Intelligence in the United States.
For details on California’s comparable AI transparency law, visit the Transparency in Frontier Artificial Intelligence Act (SB-53).
