New York Governor Kathy Hochul launched a new state office Monday to enforce safety compliance for large artificial intelligence companies, and said the state could eventually require an emergency shutoff mechanism - a "kill switch" - if AI systems pose catastrophic risk. The move places New York alongside California in building some of the strongest state-level AI guardrails in the absence of federal action.
Hochul appointed Marc Gilman, an attorney from the Department of Financial Services, to lead the new Digital Innovation, Governance, Integrity and Trust office, known as DIGIT. The office was created by the RAISE Act, which the governor signed in December 2025. The law mandates that major AI developers meet safety and transparency standards that rank among the nation's toughest restrictions on firms like OpenAI and Anthropic.
Developers must register with the state in November, with compliance enforcement starting in January. Companies will report to the DIGIT office, publish their safety protocols, and disclose any critical incidents within 72 hours or face civil penalties.
Kill switch proposal follows California's lead
Hochul said New York intends to go beyond the current legislation. "Safeguards like AI kill switches if they're deemed feasible and in the best interests of our state" could be part of future protections, she said. The remark echoes a move by California Governor Gavin Newsom, who signed an executive order Friday proposing that companies implement an emergency shutoff if AI models behave dangerously. Neither governor has detailed how such a mechanism would function technically.
Hochul criticized the Trump administration's resistance to national AI regulation, calling it "an abdication of responsibility of epic proportions." She said states must fill the regulatory void while Washington stays on the sidelines.
New office, new enforcement chief
Gilman previously served as general counsel and vice president of compliance at Theta Lake, a software company. Hochul said his background gives him insight into how major law firms advising AI companies operate. "He understands what the big lawyers at the firms will be saying," she said during a press conference. "He knows how to make sure they're staying in line."
The RAISE Act requires large developers to register with the state and comply with incident reporting and transparency standards. The law targets the largest AI firms and mirrors California's approach in setting binding safety requirements rather than voluntary guidelines. Professionals tracking state-level AI regulation can find relevant training through AI Regulatory Compliance Courses designed for regulatory affairs specialists.
Why this matters for government, legal, and communications professionals
New York's enforcement timeline is concrete: registration opens in November, and penalties for non-compliance begin in January. Legal and compliance teams at AI companies, as well as their outside counsel, need to prepare safety protocols and incident response procedures now to meet the 72-hour reporting window. Government affairs and PR professionals should expect the kill switch concept to become a recurring talking point as more states consider AI safety legislation. Policy-focused training such as AI Public Policy Courses can help teams understand the shifting regulatory landscape. The insurance sector, which underwrites technology errors and omissions policies, will also be watching how enforceable safety standards affect liability exposure for AI developers.
Your membership also unlocks:
