New York RAISE Act (Responsible AI Safety and Education Act)

Summary

New York's Responsible AI Safety and Education (RAISE) Act takes effect January 1, 2027. Codified as General Business Law Article 44-B (Sections 1420–1429), it requires developers of frontier AI models — foundation models trained with more than 10^26 operations — to publish transparency reports and report critical safety incidents within 72 hours, and requires large frontier developers (over $500 million in annual revenue with affiliates) to publish a frontier AI framework, file a disclosure statement with a new office in the Department of Financial Services, and pay assessments. The Attorney General may seek civil penalties of up to $1 million for a first violation and $3 million for each subsequent violation.

The RAISE Act was enacted in two steps. Governor Hochul signed S6953-B/A6453-B on December 19, 2025 as Chapter 699 of the Laws of 2025, with agreed chapter amendments. Those amendments passed as S8828/A9449 (Gounardes) and were signed on March 27, 2026 as Chapter 96 of the Laws of 2026. The sponsor's memo describes the bill as "a chapter amendment to L.2025, c.699, to repeal the underlying chapter and establish a new regulatory regime": Chapter 96 repeals the Article 44-B text that Chapter 699 added and replaces it, and changes Chapter 699's effective date from the ninetieth day after enactment to January 1, 2027. The 2025 bill's compute-cost thresholds and penalty formula therefore never take effect. [Sources: S8828 text and history, S6953-B history, Governor's December 19, 2025 release, retrieved 2026-10-12.]

The replacement Article closely tracks California's SB 53 (TFAIA): the same 10^26-operation frontier-model definition, the same $500 million large-frontier-developer revenue test, and parallel framework, transparency-report and incident-reporting duties. The differences are the New York oversight office, the disclosure-statement and assessment regime, and higher maximum penalties.

Compliance timeline

Who is covered

• Frontier model — a foundation model trained using more than 10^26 integer or floating-point operations, counting the original training run and any later fine-tuning, reinforcement learning, or other material modification (§ 1420(9)).
• Frontier developer — a person who has trained, or initiated the training of, a frontier model (§ 1420(8)).
• Large frontier developer — a frontier developer that, together with its affiliates, had annual gross revenues above $500,000,000 in the preceding calendar year (§ 1420(10)).

The Article applies to frontier models developed, deployed, or operating in whole or in part in New York (§ 1425). Accredited New York colleges and universities doing academic AI research, and the Empire AI consortium, are excepted (§ 1426).

Developer duties

A catastrophic risk is a foreseeable and material risk that a frontier model will materially contribute to the death of, or serious injury to, more than 50 people, or more than $1 billion in property damage, from a single incident in which the model provides expert-level help with a chemical, biological, radiological, or nuclear weapon, engages without meaningful human oversight in a cyberattack or conduct that would be murder, assault, extortion, or theft if done by a person, or evades its developer's or user's control (§ 1420(3)).

Oversight office and fees

The Article assigns implementation to an office within the Department of Financial Services that reports to the Superintendent (§ 1420(16)) and has rulemaking authority (§ 1429). On September 21, 2026 the Governor named it the Office of Digital Innovation, Governance, Integrity and Trust (DIGIT) and appointed Marc Gilman as Deputy Director for the RAISE Act. The office runs the incident-reporting mechanism, which members of the public may also use, and from January 1, 2028 publishes an annual report. Failing to file a disclosure statement, or leaving false information uncorrected, can bring a civil penalty of $1,000 for each day after notice and a hearing, plus the unpaid assessments (§ 1428).

Enforcement and penalties

The Attorney General may bring a civil action for a penalty of up to $1,000,000 for a first violation and up to $3,000,000 for each subsequent violation, scaled to severity, when a large frontier developer fails to publish or transmit a required document, makes a prohibited false or misleading statement, fails to report an incident, or fails to comply with its own frontier AI framework (§ 1427(1)). The Article expressly creates no private right of action (§ 1427(2)).

Frequently asked questions

When does the New York RAISE Act take effect?

January 1, 2027. Chapter 96 of the Laws of 2026 (S8828, signed March 27, 2026) changed the original Chapter 699 effective date, which would have been the ninetieth day after enactment, to January 1, 2027.

Who must comply with the RAISE Act?

Developers that have trained, or begun training, a foundation model with more than 10^26 operations carry the transparency-report and incident-reporting duties. Large frontier developers, those with more than $500 million in annual gross revenue including affiliates, also carry the frontier AI framework, quarterly internal-use reporting, and disclosure-statement duties.

How fast must a critical safety incident be reported under the RAISE Act?

Within 72 hours of determining that a critical safety incident occurred, or of learning facts sufficient to establish a reasonable belief that one occurred, to the DFS office. If the incident poses an imminent risk of death or serious physical injury, the developer must also disclose it within 24 hours to an appropriate law-enforcement or public-safety authority.

What are the penalties under the New York RAISE Act?

The Attorney General may seek up to $1 million for a first violation and up to $3 million for each subsequent violation. Separately, the DFS office may impose $1,000 per day for failing to file the required disclosure statement. There is no private right of action.