California is positioning itself to lead on AI oversight, with Governor Gavin Newsom issuing an executive order that opens the door to mandating a kill switch for frontier models. The order, signed on Friday, directs the state to convene a group of experts tasked with delivering recommendations within two months on how to strengthen AI safety provisions in state law.
The panel has been asked to weigh several measures. These include requiring AI companies to embed independent verification groups onsite for regular audits, and making transparency reports and risk assessments subject to the standards of independent auditors. The experts will also consider the creation of a kill switch that is routinely verified as effective, along with rules that would compel companies to report so-called loss-of-control incidents as critical safety events.
Building on recent legislation
The executive order also instructs a state agency to accelerate implementation of two laws Newsom recently signed. One establishes a framework for independent verifiers to assess AI safety, while the other creates a state registry of AI auditors. Together, the measures form the basis of a wider system of accountability that California intends to expand.
Newsom has signalled that further action may follow. He recently raised the possibility of calling California lawmakers in for a special legislative session focused on AI, alongside potential additional executive action. He has framed the current order as a template for federal policymakers, urging Congress and President Donald Trump to review and adopt the state’s framework, or to treat its regulation as a floor rather than a ceiling.
Federal inaction and the wider debate
The move comes amid renewed debate over AI risk. Discussions of various bills have gathered pace in Congress following a warning from a former Anthropic researcher of a greater than 10 percent chance that AI could kill all humans within a decade. Prospects for federal legislation remain limited, however, with the House in recess until November and the approaching midterm elections reducing the likelihood of major action in the Senate.
President Trump has taken a markedly different stance, arguing that the only guardrails AI requires are a strong and smart president, and asserting that the United States already has one.
In a statement, Newsom criticised what he described as the federal government’s failure to create meaningful AI oversight or accountability, noting that AI chief executives themselves have called for regulation. He said the state would not wait to act and would move with what he described as urgent velocity on responsible oversight.
Source
Image: theverge.com
