LawSnap

NY Gov. Hochul Signs Final RAISE Act Amendments for Frontier AI on March 27, 2026

Published April 6, 2026

Score

Why it matters

On March 27, 2026, New York Governor Kathy Hochul signed chapter amendments finalizing the Responsible AI Safety and Education (RAISE) Act, regulating developers of frontier AI models—defined as models trained with over (10^{26}) FLOPs and compute costs exceeding $100 million, including those via knowledge distillation.[1][3][8] The law takes effect January 1, 2027, applying to developers with annual revenues over $500 million operating in New York, requiring safety protocols, 72-hour incident reporting, transparency reports, annual frameworks, and assessments by a new DFS office; accredited universities are exempt.[1][3][5][8]

Key players include Governor Kathy Hochul, sponsors State Senator Andrew Gournardes and Assemblymember Bores, the New York Department of Financial Services (DFS) for oversight, and the New York Attorney General for enforcement (penalties capped at $1M first offense, $3M repeat).[3][5][8][10] It aligns closely with California's TFAIA (SB-53), adopting identical revenue thresholds after negotiations to harmonize rules across tech states.[1][2][3][7]

The RAISE Act passed initially in June 2025, was signed December 19, 2025, with amendments introduced January 6, 2026, passed March 11, and finalized March 27 amid federal inaction post-White House EO on minimal regulation and DOJ AI Task Force formation.[3][4][9] Newsworthy now as it cements New York's "nation-leading" AI safety standard just before the April 2026 effective amendments, preempting federal patchwork risks while easing multi-state compliance for developers like those in CA-NY hubs.[3][4][5][7]