New York's Responsible AI Safety and Education Act (RAISE Act) for Large Developers
Issued by
New York State Legislature
- September 30, 2026 · Correction — Updated the entry to the enacted and amended law: revenue threshold, published framework, incident reporting, DFS office, and penalties (fact-check finding). (Cody Maxwell)
- October 1, 2026 · Correction — Rewrote the practical steps, newsletter hook, search description, and audience fields to match the entry's corrected content. (Cody Maxwell)
New York's RAISE Act was signed on 19 December 2025 and amended on 27 March 2026. From 1 January 2027, frontier model developers with over $500 million in annual revenue must publish a frontier AI framework, issue transparency reports, and report critical safety incidents. A new office in the Department of Financial Services oversees them.
Applies To
Overview
Governor Hochul signed the RAISE Act on 19 December 2025, and a chapter amendment finalizing its text was signed on 27 March 2026. It takes effect on 1 January 2027. It applies to large frontier developers: developers of frontier models, trained with more than 10^26 operations, with over $500 million in annual revenue. The amendment replaced the original compute-cost thresholds and the original annual third-party audit requirement. Covered developers must publish a frontier AI framework describing their safety practices, including any use of third parties for assessments. They must publish a transparency report before deploying a new frontier model. Critical safety incidents must be reported within 72 hours, or within 24 hours to law enforcement or public safety agencies when danger is imminent. A new oversight office in the Department of Financial Services receives disclosure statements, renewed every two years or on a change of ownership. It can fine developers $1,000 a day for failing to file. The Attorney General enforces the Act, with penalties of up to $1 million for a first violation and $3 million for later ones.
Key Requirements
- •Applies to developers of frontier models (over 10^26 operations) with more than $500 million in annual revenue.
- •Publish a frontier AI framework describing safety practices, including any third-party assessments.
- •Publish a transparency report before deploying a new frontier model.
- •Report critical safety incidents within 72 hours, or 24 hours where danger is imminent.
- •File a disclosure statement with the DFS oversight office, renewed every two years; late filing costs $1,000 a day.
- •The Attorney General enforces, with penalties up to $1 million for a first violation and $3 million for later ones.
What Your Organization Must Do
- →Check whether your models exceed 10^26 training operations and your revenue tops $500 million.
- →Draft and publish a frontier AI framework describing safety practices and any third-party assessments.
- →Build a transparency report process that completes before each new frontier model is deployed.
- →Set up incident triage that meets the 72 hour reporting clock and the 24 hour imminent danger clock.
- →File the disclosure statement with the Department of Financial Services office and diary the two year renewal.
- →Brief leadership on penalties of up to $1 million first time and $3 million later.