Rules / California

Transparency in Frontier Artificial Intelligence Act (SB 53)

In forceLawBinding

Requires developers of frontier models (trained with more than 10^26 operations) to publish transparency reports at release, and 'large frontier developers' (over $500M annual revenue) to publish and follow a frontier AI framework on catastrophic risk and send quarterly risk summaries to Cal OES. Critical safety incidents must be reported to Cal OES within 15 days (24 hours if there is imminent risk of death or serious injury), and whistleblowers are protected.

Why it matters

The first binding US frontier-model safety law; it is the template copied by New York's RAISE Act and Illinois SB 315 and a prime target of federal preemption efforts.

What it requires

Frontier model safetyTransparency to usersIncident reportingPenalties

Penalties

Civil penalty up to $1,000,000 per violation, enforced by the California Attorney General

Cite this record

Transparency in Frontier Artificial Intelligence Act (SB 53). California Legislature / Governor. Status: In force. wheresthe.ai, https://wheresthe.ai/rule/us-ca-sb53-tfaia-2025/ (verified 4 Oct 2026).

Line of Thought

Developments connected to this rule

What has happened that this rule shapes, or that shapes it. Pick one to keep following the thread.

More from California

  • AB 2013
    LawCalifornia Legislature / GovernorBinding
    In forceEffective 1 Jan 2026
  • CCPA ADMT regulations
    RegulationCalifornia Privacy Protection Agency (CPPA)Binding
    Partly in forceEffective 1 Jan 2026
  • SB 243
    LawCalifornia Legislature / GovernorBinding
    In forceEffective 1 Jan 2026
  • SB 1119 / Adam's Law
    LawCalifornia Legislature / GovernorBinding
    Enacted, not yet in forceApplies 1 Jul 2027
  • CA AI Transparency Act
    LawCalifornia Legislature / GovernorBinding
    Partly in forceEffective 2 Aug 2026
  • SB 947
    LawCalifornia Legislature / GovernorBinding
    Enacted, not yet in forceApplies 1 Jul 2027