EFF Response to Governor Newsom’s AI Kill Switch Executive Order

Accelerating Independent Audits and Oversight Frameworks

Under the executive order, Governor Newsom directed the Government Operations Agency to expedite implementation timelines for Senate Bill 813 and Assembly Bill 1405, according to the official state release. Senate Bill 813 established a framework for independent verification organizations to assess AI systems and models for safety risks. Meanwhile, Assembly Bill 1405 created a state registry for AI auditors to enforce compliance, transparency, and integrity standards. The new directive instructs the Government Operations Agency, alongside the Governor’s Office of Emergency Services, to convene national experts. These experts will help draft recommendations requiring frontier AI companies to embed a designated independent verification organization onsite in their laboratories. These onsite auditors will conduct regular evaluations and verify that mandatory safety frameworks, transparency reports, and risk assessments comply with state law.

Developing an AI Kill Switch and Expanding Incident Reporting

The executive order also pushes state regulators to advance the creation of an AI kill switch for frontier models, with efficacy verified continuously by an independent organization, according to government records. Furthermore, the administration will update the definitions of critical safety incidents under Senate Bill 53 from 2025—known as the Transparency in Frontier Artificial Intelligence Act—to formally include loss-of-control incidents like the Hugging Face attack. These measures follow federal inaction on AI security standards. According to the state’s announcement, the federal government maintains no law requiring AI companies to report dangerous incidents, and President Trump has rejected calls for new regulations to protect public safety. Governor Newsom’s administration is calling on Congress and President Trump to review California’s framework and adopt it as a baseline for national standards.

Stakeholder Reactions and Practical Concerns

While state officials promote the executive order as a model for national safety, advocacy groups have raised specific operational concerns. According to a statement released by the Electronic Frontier Foundation, the organization welcomes the dialogue regarding potential AI harms but emphasizes that the most immediate risks involve biased algorithmic decision-making in employment and government benefits, automated surveillance systems such as Flock cameras, and artificially inflated personalized pricing. The Electronic Frontier Foundation also cautioned that government-controlled kill switches carry inherent risks of being weaponized against protected speech, citing retaliatory actions taken against Anthropic by the Trump administration earlier in the year.

As the Government Operations Agency prepares its recommendations for the governor, we urge leaders to also realize that the effectiveness of kill switches in advanced AI systems remains an area of active research.

Electronic Frontier Foundation The organization urged state regulators to ensure that technical cybersecurity mandates for AI laboratories remain precise and practical, while also ensuring that independent investigation programs are made accessible to smaller developers.