An overview of AI safety in California state politics
California State Politics AI Safety Overview
The author Felix Moses reviews California AI safety-related legislation and considers that the substantive regulation currently in effect is mainly SB 53, which requires frontier developers to publish safety frameworks, report serious incidents, and protect whistleblowers. SB 813 establishes voluntary independent verification organizations (IVOs) certified by the state government; audit results are relevant to liability in litigation but are not conclusive, and laboratories are not required to accept audits; certification standards are scheduled to be issued before May 1, 2027; AB 1405, starting in 2029, imposes mandatory registration on AI audit institutions. Executive Order N-9-26, signed on September 18, 2026, requires state agencies to accelerate implementation of audit laws and proposes recommendations on mandatory independent verification, resident auditors, kill switches, and expanded incident reporting. The author lists binding safety requirements and third-party audit transparency as high priority, kill-switch authority as medium priority, and mentions SB 1047, which was vetoed by Newsom in 2024, while also listing organizations working in this field such as Encode AI, Fathom, and Secure AI Project.