Particle.news

Download on the App Store

California Enacts First-in-Nation Frontier AI Transparency Law

California sets a disclosure-focused rule for frontier AI developers that supporters say could guide national policy.

Overview

  • Gov. Gavin Newsom signed SB 53 on Sept. 29, 2025, with the law taking effect Jan. 1, 2026.
  • The measure targets large frontier developers, defined by training runs exceeding 10^26 operations and annual revenue over $500 million.
  • Companies must publish a public frontier AI framework and pre-deployment transparency reports detailing capabilities, risks, mitigations, and any third-party evaluations.
  • Developers must report critical safety incidents to the state within 15 days and notify appropriate authorities within 24 hours if imminent harm is identified, with added whistleblower protections.
  • The California Attorney General can seek civil penalties up to $1 million per violation, and the law establishes the University of California–operated CalCompute consortium with a framework due by Jan. 1, 2027, as industry reactions range from support by OpenAI, Meta, and Anthropic to opposition from trade groups.