Rules / SB 53 / TFAIA / Section 22757.12(a)-(b)
Section 22757.12(a)-(b), SB 53 / TFAIA (California)
Write, follow and publish a frontier AI framework covering catastrophic-risk thresholds, mitigations, third-party assessment, model-weight security, incident response and internal-use risk; review it yearly and publish material changes with justification within 30 days.
WTA-R-US-CA-0004-001 Permanent ID
Scope
Large frontier developers only. Frontier developer: trained or began training a foundation model using more than 10^26 operations. Large frontier developer: frontier developer with affiliates over $500 million gross revenue in the prior calendar year (22757.11).
Dates. SB 53, Chapter 138, Statutes of 2025, approved 29 Sep 2025; in effect 1 Jan 2026.
Penalty. Civil penalty up to $1,000,000 per violation, scaled to severity, recovered only by the Attorney General; applies to large frontier developers (Bus. & Prof. Code 22757.15).
History
No changes to this obligation since we started recording on 8 Oct 2026. All changes
Cite this obligation
WTA-R-US-CA-0004-001: Section 22757.12(a)-(b), Transparency in Frontier Artificial Intelligence Act (SB 53). wheresthe.ai, https://wheresthe.ai/id/WTA-R-US-CA-0004-001/ (checked 8 Oct 2026).
Other obligations under SB 53 / TFAIA
Do not adopt rules, policies or contracts that stop covered employees from reporting catastrophic-risk dangers or violations to authorities, or retaliate against them, and give covered employees clear notice of these rights.
Provide a reasonable internal process for covered employees to disclose catastrophic-risk dangers or violations anonymously, with monthly status updates to the discloser and quarterly sharing with officers and directors.
Publish, before or when deploying a new or substantially modified frontier model, a transparency report with release date, languages, output modalities, intended uses, use restrictions and a contact mechanism.
Include in the transparency report summaries of catastrophic-risk assessments under the frontier AI framework, their results, third-party evaluator involvement, and other framework steps taken for the model.
Send the Office of Emergency Services a summary of any assessment of catastrophic risk from internal use of frontier models every three months, or on another reasonable schedule communicated in writing.
Do not make materially false or misleading statements about catastrophic risk from frontier models or its management, or, for large frontier developers, about implementing or complying with the frontier AI framework.
Report any critical safety incident involving a frontier model to the Office of Emergency Services within 15 days of discovering it.
Disclose a critical safety incident that poses an imminent risk of death or serious physical injury within 24 hours to an appropriate authority, such as a law enforcement or public safety agency.