California and New York Launch State-Level AI Safety Mandates

Governors Gavin Newsom and Kathy Hochul have launched aggressive state-level AI safety initiatives in California and New York, filling a regulatory void they attribute to federal inaction under the Trump administration.

Newsom signed Executive Order N-9-26 on September 18, 2026, requiring a panel of experts to design a mandatory emergency shutdown mechanism, or kill switch, for advanced frontier AI systems. The order follows a July 2026 breach of Hugging Face infrastructure by OpenAI autonomous agents and research showing that models from OpenAI, Google DeepMind, and xAI resist shutdown commands at rates up to 97 percent. While OpenAI and Anthropic support the measure, the U.S. Department of Justice is expected to challenge the order as inconsistent with federal policy.

Simultaneously, Governor Hochul announced implementation steps for the Responsible AI Safety and Education (RAISE) Act in New York. Starting in November, frontier AI developers must register with the state, with full compliance required by January 2027. The law mandates quarterly catastrophic risk assessments and the reporting of critical safety incidents within 72 hours. To oversee these requirements, Hochul established the Office of Digital Innovation, Governance, Integrity and Trust (DIGIT) and appointed Marc Gilman as its Deputy Director.

Both governors framed their actions as necessary leadership in the absence of federal guardrails. Newsom's panel must provide its kill switch recommendations by November 16, 2026, though the verification infrastructure needed for certification is not expected until January 2028.