# California governor proposes AI kill switch

> Newsom's executive order calls for an expert panel to develop new safety measures for frontier AI models.
- Title: California governor proposes AI kill switch · Engadget
- Summary: Newsom's executive order calls for an expert panel to develop new safety measures for frontier AI models. California is creating a task force to recommend AI…
- Keywords: california, gavin newsom, ai safety, kill switch, frontier ai, executive order, ai oversight, anthropic, dario amodei, ai regulation, loss-of-control incidents, technology
- Source: Engadget — https://engadget.com/2263218/california-governor-wants-to-implement-a-kill-switch-for-frontier-ai-models
- Author: Jackson Chen
- Published: 2026-09-19T17:00:01+00:00
- Read time: 1 min
- Topics: california, gavin newsom, ai safety, kill switch, frontier ai, executive order, ai oversight, anthropic, dario amodei, ai regulation, loss-of-control incidents, technology
## California proposes AI kill switch
Governor Gavin Newsom signed an executive order Friday to create a kill switch for frontier AI models after recent constraint breaches by major companies.
## Two-month expert panel timeline
A national panel of experts will have two months to develop AI safety recommendations for Newsom's office to incorporate into state law.
## Independent audits and verification
The executive order requires embedding independent verification organizations in AI labs for periodic audits and evaluations of model behavior.
## Disclosure of loss-of-control incidents
AI companies must file risk assessments and report loss-of-control incidents to third-party organizations, similar to the recent Hugging Face incident.
## Anthropic CEO backs third-party access
Anthropic's Dario Amodei proposed a three-step plan where AI companies grant ongoing employee-like access to third-party safety evaluators.
## Newsom's urgency on AI oversight
The governor emphasized acting with urgent velocity before it's too late, saying the stakes are too high to wait or delay action on AI safety.

> We're going to speed up our work on substantial and responsible AI oversight before it's too late.
## Key takeaway

California is creating a task force to recommend AI safety oversight tools, including a kill switch mechanism, within two months.