Particle.news

Download on the App Store

California Enacts First-of-Its-Kind Frontier AI Transparency Law

It applies to the biggest AI labs through compute and revenue thresholds, focusing penalties on failures to disclose.

Overview

  • Developers of frontier models must publish a public safety framework and pre-deployment transparency reports with risk assessments and any third-party evaluations.
  • Critical safety incidents must be reported to California’s Office of Emergency Services within 15 days, or within 24 hours if there is an imminent risk of death or serious injury.
  • The statute protects covered employees who raise catastrophic-risk concerns, enabling anonymous internal reports and court injunctions with attorneys’ fees for retaliation.
  • Enforcement rests with the Attorney General with civil penalties up to $1 million per violation, and coverage is limited to models using more than 10^26 operations and firms with over $500 million in revenue.
  • Compared with the vetoed SB 1047, the new law drops a mandated “kill switch,” emphasizes disclosures over liability for harms, establishes the CalCompute planning consortium, and takes effect January 1, 2026.