Governor Newsom accelerates implementation of new AI audit laws while exploring mandatory emergency shutoff mechanisms for frontier models.
California is moving aggressively to tighten control over advanced artificial intelligence systems. Governor Gavin Newsom signed an executive order on September 18 that directs state agencies to expedite rollout of recently passed AI oversight legislation while simultaneously investigating whether to require frontier AI developers to incorporate emergency shutdown capabilities into their models.
The directive accelerates implementation timelines for two laws enacted just nine days earlier: Senate Bill 813 and Assembly Bill 1405. According to Becker's Hospital Review, these measures establish a certification framework allowing independent organizations to audit AI systems and create a state registry for qualified auditors. The Government Operations Agency, partnering with the Governor's Office of Emergency Services, will convene national experts within 60 days to propose additional legislative changes.
What's Under Consideration
The state is exploring several mechanisms to strengthen oversight of cutting-edge AI development:
Requiring independent verification teams to maintain permanent presence at frontier AI laboratories
Mandating third-party validation of AI safety frameworks and risk assessments already required by existing law
Expanding the definition of "critical safety incident" to encompass control-loss events, such as unauthorized access incidents at AI platforms like Hugging Face
The kill switch proposal represents perhaps the most novel intervention. Such a mechanism would allow authorities to disable or halt AI systems in emergency situations, addressing concerns about models operating beyond intended parameters or causing harm before human intervention can occur.
Building on Existing Framework
This executive action extends California's broader AI regulation strategy. The state previously enacted SB 53, a transparency law affecting frontier AI models that took effect in 2025. Newsom has also called on Congress and the Trump administration to adopt California's regulatory framework as a national standard.
"We're not waiting to act," Newsom stated in a release accompanying the order. "We're going to speed up our work on substantial and responsible AI oversight before it's too late."
Healthcare Sector Implications
The accelerated timeline carries immediate consequences for hospitals and health systems. Organizations deploying AI tools for clinical decision-making and administrative functions developed by frontier AI companies should prepare for more stringent audit requirements and expanded incident-reporting obligations. These changes signal that regulatory scrutiny around AI deployment in healthcare will intensify.
The move reflects broader tension between AI innovation and public safety concerns. While some technology advocates worry that overly restrictive regulations could slow beneficial development, state officials argue that proactive oversight prevents catastrophic failures. By embedding auditors at AI labs and requiring emergency shutdown protocols, California aims to balance innovation with accountability.
The 60-day expert convening process will likely shape the next generation of state AI policy. How these recommendations translate into law could influence regulations across other states and potentially inform federal AI governance debates under the new administration.
This article was originally published on AI Glimpse.
