New York RAISE Act

New York, United StatesS.6953-B / A.6453-B (Ch. 699 of 2025), amended by S.8828 / A.9449 (Ch. 96 of 2026)Updated September 21, 2026

100%became law
Bad for your rightsLaw
  1. Introduced
  2. Passed House
  3. Passed Senate
  4. Became Law

Signed Dec 19, 2025; rewritten by chapter amendment Mar 27, 2026; takes effect Jan 1, 2027

Makes the largest AI companies file a disclosure statement with a new state office inside New York's Department of Financial Services, pay its costs, publish risk frameworks, and report "critical safety incidents" within 72 hours, or face fines of up to $3 million.

Latest update

Hochul, AG Letitia James and the sponsors announced the next steps for the law, including developer registration and talk of exploring AI "kill switches"

The bill

Official bill record

Key facts

Where
New York, United States
Bill number
S.6953-B / A.6453-B (Ch. 699 of 2025), amended by S.8828 / A.9449 (Ch. 96 of 2026)
Status
Law
Introduced
March 5, 2025
Threat
Medium
Last update
September 21, 2026

Topics

Share this bill

What it does

The RAISE Act adds Article 44-B to New York's General Business Law. The version that takes effect is the March 2026 chapter amendment (S.8828, Chapter 96 of 2026), which repealed and replaced the 2025 text. It covers "frontier models," meaning AI foundation models trained with more than 10^26 operations, including later fine-tuning, that are developed, deployed or operating in whole or in part in New York. The heaviest duties fall on "large frontier developers" with more than $500 million in annual revenue, counting affiliates. These companies must write, follow and publish a "frontier AI framework" describing how they assess and mitigate "catastrophic risk" (more than 50 deaths or serious injuries, or over $1 billion in damage, tied to weapons help, autonomous crimes or cyberattacks, or a model evading control). They must review it yearly and publish any material change within 30 days. Before or when it deploys a new or substantially modified model, a developer must publish a transparency report, and large developers must add summaries of their risk assessments and third-party testing. Developers must report "critical safety incidents" to a new office inside the Department of Financial Services within 72 hours, and within 24 hours to law enforcement when there is an imminent risk of death or serious injury. Large developers must also send the office confidential summaries of risks from internal model use every three months. No large frontier developer may develop, deploy or operate a covered model in New York without a current disclosure statement on file with the office, renewed every two years or sooner after a transfer or material change, listing its offices, contacts and owners of 5 percent or more (50 percent for public companies), and without paying a pro rata assessment to fund the office. The office can take incident reports from the public, share reports with other agencies including the Attorney General, send an annual report to the governor and legislative leaders from 2028, and adopt rules, including "additional reporting or publication requirements." The Attorney General enforces the law with civil penalties of up to $1 million for a first violation and up to $3 million for later ones. The office can separately fine $1,000 per day for a missing or false disclosure statement. There is no private right of action. Academic research at New York colleges and the Empire AI consortium are exempt.

What’s at stake

The law turns the right to build and run the most capable AI models in New York into something a company must keep current on file with a financial regulator and pay for. The current duties are mostly paperwork and disclosure, and the thresholds reach only a handful of very large companies. The office's open rulemaking power, its right to demand more reporting, and a filing requirement that ties operating in the state to the regulator create a standing channel of state oversight over general-purpose AI systems that millions of people use to write, search and speak. State officials have already said they may "explore" further measures such as AI "kill switches," which are not in the law today.

Our take

The RAISE Act sets up a state office that the biggest AI developers must file with, report to and pay for before they can build or run their models in New York. Today's text is narrower than the 2025 version, with lower fines and disclosure duties in place of a ban on deploying risky models. The structure is the concern: a regulator with open-ended rulemaking power over general-purpose software that people use to write and communicate, and officials already talking about "kill switches." Once the filing and reporting machinery exists, adding demands requires only a rule.

Timeline

  1. Law takes effectUpcoming

  2. Hochul, AG Letitia James and the sponsors announced the next steps for the law, including developer registration and talk of exploring AI "kill switches"Latest

  3. Chapter amendment signed as Chapter 96 of 2026; RAISE Act text replaced and effective date set to Jan 1, 2027

  4. The chapter amendment cut maximum penalties from $10 million/$30 million to $1 million/$3 million, dropped the ban on deploying a model with "unreasonable risk of critical harm," moved reporting from the Attorney General and the homeland security division to a new office in the Department of Financial Services, and added the filing and fee requirement and rulemaking power

  5. S.8828 passed the Assembly 140-2

  6. S.8828 passed the Senate

  7. Chapter amendment S.8828 introduced by Sen. Gounardes

  8. Signed as Chapter 699 of 2025 with approval memo 76, on agreement to a chapter amendment

  9. Delivered to Gov. Kathy Hochul

  10. Passed the Senate; passed the Assembly 119-22

  11. Amended in both chambers to the final S.6953-B / A.6453-B versions

  12. S.6953 introduced by Sen. Gounardes; referred to Senate Internet and Technology Committee

  13. A.6453 introduced by Assemblymember Bores; referred to Assembly Science and Technology Committee

The record

Sponsors

Sen. Andrew Gounardes (D), Assemblymember Alex Bores (D)

Committee

Senate Internet and Technology Committee, Senate Rules Committee, Assembly Science and Technology Committee, Assembly Codes Committee, Assembly Ways and Means Committee, Assembly Rules Committee

Last updated September 21, 2026