The law layer · New York, United States · US state law
New York: the RAISE Act
Responsible AI Safety and Education Act, S6953 · signed by Governor Kathy Hochul on 19 December 2025 · effective 1 January 2027
Page verified 5 August 2026. Signing date and effective date confirmed against the Governor’s office and the New York Senate record. Obligation detail is summarised, and we say so.
The second American state to legislate on frontier AI, and the one that shows the pattern was not a California anomaly. Signed three months after SB 53, in a different state, by a different governor, requiring substantially the same disclosure.
The essentials
- Signed19 December 2025, by Governor Kathy Hochul.
- Effective1 January 2027.
- BillS6953, the Responsible AI Safety and Education Act.
- DutyLarge frontier developers must establish and publish safety protocols.
- ApproachTransparency and safety requirements applied to frontier developers and frontier models.
Where it sits
Three American states have now legislated on frontier AI within ten months, and the sequence tells you where this is going:
- Sep 2025California SB 53. Publish risk assessments and standards documentation.
- Dec 2025New York RAISE Act. Publish safety protocols.
- Jul 2026Illinois SB 315. Publish, and be audited by an independent third party.
Each state added a turn of the screw. The first two ask developers to say what they do. The third asks someone else to verify it, which is the moment publication stops being sufficient.
The federal question
All three state laws were enacted against an active debate about federal preemption of state AI regulation. That debate is unresolved, and anyone planning around these statutes should track it. We record the state laws because they are enacted and dated; we do not speculate about whether Congress will displace them.
What to do about it
This is an estate problem, not a document problem.
A published safety protocol is a promise about how a system will be run. What turns it into evidence is a record of the individual decisions taken under it. Regulayer records the decision as it is taken, outside the system being governed, content-free, and verifiable by anyone holding the file with public mathematics, without a call to us.
Regulayer for enterpriseDrift, what changed and when →
For an individual practitioner, the same engine is $349 a month with a free week: start sealing. See also the law layer and every jurisdiction we track.
Sources
- Office of the Governor, signing announcement
- New York State Senate, S6953
- No section-by-section block: obligations summarised from official announcements and counsel analysis rather than a section read of the enacted text.
Related
Checked 5 August 2026. Information, not legal advice.
Common questions
What is the New York RAISE Act?
The Responsible AI Safety and Education Act, signed by Governor Kathy Hochul on 19 December 2025. It makes New York the second US state after California to enact comprehensive legislation regulating the safety and governance of frontier AI models.
When does the RAISE Act take effect?
1 January 2027.
What does the RAISE Act require?
It shapes AI safety through transparency and safety requirements applicable to frontier developers and frontier models. Large frontier developers must establish and publish safety protocols.
Does the RAISE Act affect ordinary New York businesses?
Not directly. Like California SB 53 and Illinois SB 315, it reaches large frontier model developers rather than the businesses that deploy their models. Its indirect effect is that the models New York businesses use will come with published safety protocols.
Information, not legal advice, and not advice on the law of any jurisdiction. Every entry is verified against the issuing body’s own document; where a source is reporting rather than the document, we say so.