California Transparency in Frontier Artificial Intelligence Act

Section 22757.10 names the chapter the Transparency in Frontier Artificial Intelligence Act. It was added by Statutes 2025, Chapter 138 (SB 53), effective 1 January 2026.

The Act is California state law. It creates public-framework, transparency-report, incident-reporting and anti-misstatement duties for covered frontier developers, with elevated duties for large frontier developers. It is not a federal predeployment licensing statute.

Section 22757.12(a) requires a large frontier developer to write, implement, comply with, and clearly and conspicuously publish on its website a frontier AI framework covering its frontier models. The framework must describe approaches to standards incorporation; catastrophic-risk thresholds and assessments; mitigations; deployment and extensive internal-use review; third-party assessment; framework updates and substantial-modification triggers; cybersecurity of unreleased weights; critical safety incident response; internal governance; and catastrophic risk from internal use, including circumvention of oversight.

Section 22757.12(b) requires at least annual review and update of the framework. Material modifications must be published with justification within 30 days.

Section 22757.12(c) requires frontier developers, before or concurrently with deploying a new or substantially modified frontier model, to publish a transparency report covering website, contact mechanism, release date, languages, modalities, intended uses and use restrictions. Large frontier developers must also summarise catastrophic-risk assessments, results, third-party involvement and other framework steps. Publishing the same information inside a larger system or model card counts as compliance.

Section 22757.12(d) requires large frontier developers to transmit catastrophic-risk assessment summaries from internal use to the Office of Emergency Services every three months, or on another schedule communicated in writing to OES.

Section 22757.12(e) prohibits materially false or misleading statements about catastrophic risk or, for large frontier developers, about framework implementation, with a good-faith reasonableness defence. Subdivision (f) allows limited redactions for trade secrets, cybersecurity, public safety, national security or conflicting law, with character-and-justification descriptions and five-year retention of unredacted material.

• Large frontier developer · publish and follow a frontier AI framework (§ 22757.12(a)–(b)).
• Frontier developer · publish transparency reports on new or substantially modified models (§ 22757.12(c)).
• Large frontier developer · quarterly (or agreed) OES summaries of internal-use catastrophic-risk assessments (§ 22757.12(d)).
• Frontier developer · report critical safety incidents to OES within 15 days; 24 hours if imminent death or serious injury risk (§ 22757.13).
• Attorney General only · civil penalty up to $1,000,000 per violation for listed failures (§ 22757.15).

Section 22757.13 requires OES to create public and developer reporting channels for critical safety incidents, and a confidential channel for large frontier developers’ internal-use catastrophic-risk summaries. Frontier developers must report critical safety incidents to OES within 15 days of discovery, or within 24 hours to an appropriate authority if the incident poses an imminent risk of death or serious physical injury.

OES may designate substantially equivalent federal incident-reporting laws or guidance. A developer that declares intent to comply via those federal standards is deemed compliant with § 22757.13 to that extent until the declaration or designation is revoked. Failure to meet the designated federal standard then counts as a chapter violation.

From 1 January 2027, OES must produce annual anonymised aggregated reports on reviewed critical safety incidents. Section 22757.14 requires the Department of Technology, from 2027, to recommend annual updates to the definitions of frontier model, frontier developer and large frontier developer. The Attorney General produces a parallel anonymised report on covered-employee reports under Labor Code Chapter 5.1.

Section 22757.15 authorises civil penalties of up to one million dollars per violation, recoverable only in an Attorney General action, for failures to publish or transmit required documents, misstatements under § 22757.12(e), incident-reporting failures, or failure to comply with the developer’s own frontier AI framework. Section 22757.16 states that loss of equity value is not property damage for chapter purposes.

The Act is a state transparency, framework-publication and incident-reporting regime. It does not create a California predeployment licence or a compute-threshold test inside the framework duty paragraph itself; those thresholds live in the definitions section and in later Department of Technology recommendations.

Federal preemption or complementary federal rules are outside the indexed chapter text. The federal-alignment pathway in § 22757.13(h)–(j) only deems compliance with California incident reporting when OES designates equivalent federal standards and the developer elects that path.

Coverage depends on statutory definitions of frontier model, frontier developer and large frontier developer in § 22757.11. From 2027 those definitions are subject to annual Department of Technology recommendations.

This page does not audit whether any named company has published a compliant framework or filed required reports.

The following actors are linked to this vehicle in the graph. Their presence here records institutional role on the page, not an independent finding that each actor has completed a compliance duty.

• California Legislature (legislature): State legislature that enacted the Transparency in Frontier Artificial Intelligence Act (SB 53).
• Sen. Scott Wiener (legislator): California senator and author of SB 53, the Transparency in Frontier Artificial Intelligence Act.
• California Office of Emergency Services (government): California agency designated to receive frontier catastrophic-risk summaries and critical safety incident reports under SB 53.

Each block below is an explicit evidence row: a claim that points to one supporting passage. Read the quote and locator before treating the paraphrase as settled.

• California requires large frontier developers to publish a frontier AI framework. The duty is a public-framework and transparency obligation for large frontier developers, not a federal predeployment licensing regime. Operational note: It creates a state-level vehicle for catastrophic-risk assessment disclosure and annual framework review.
• California’s frontier AI transparency chapter took effect on 1 January 2026. Effective-date language distinguishes enactment in September 2025 from operative duties beginning in January 2026. Operational note: It supplies the legal date needed to treat the framework and reporting duties as current California law.

Dated events attached to this vehicle. Each entry is a graph event with its own update page when present.

• 2025-09-29 · publication · California SB 53 chaptered into law. Chaptering fixed the statutory text that later took effect on 1 January 2026. Consequence recorded: Created California’s frontier-developer transparency statute.
• 2026-01-01 · effective-date · California frontier AI transparency duties take effect. The effective date moved framework publication and related duties from enacted text into operative state law. Consequence recorded: Started the operative period for large frontier developer transparency duties.

A vehicle is an instrument. Asks are the outcomes the instrument is treated as advancing in the AI Tribune graph.

• Require large frontier developers to publish catastrophic-risk safety frameworks Status: in-force. Require covered large frontier developers to write, implement, comply with and publish a frontier AI framework covering catastrophic-risk assessment, mitigations and related disclosures.

These documents are the originals behind the evidence rows. Prefer the document text over secondary paraphrase when the two diverge.

• California Business and Professions Code Chapter 25.1 · Transparency in Frontier Artificial Intelligence Act (Statute, 2026-01-01). Chaptered California statute requiring large frontier developers to publish frontier AI frameworks and related transparency reports.