Transparency in Frontier Artificial Intelligence Act (SB 53)

Record US-CA-05 · California Senate Bill 53 (2025-2026 Regular Session)

Issued by: California State Legislature; signed by Governor Gavin Newsom
Jurisdiction: California (US)
Level · region: Sub-national · North America
Country: United States
Published / adopted: 2025-09-29
In force: 2026-01-01
Language: English
Official identifier: California Senate Bill 53 (2025-2026 Regular Session)

Summary

California's frontier AI safety law. Large frontier developers must publish a frontier AI framework describing how they incorporate national and international standards and industry-consensus best practices, and must send the Office of Emergency Services summaries of their assessments of catastrophic risk from internal use of frontier models. The Office runs a mechanism for frontier developers and the public to report critical safety incidents. The Act protects whistleblowers who disclose significant risks posed by frontier models, creates a civil penalty enforced by the Attorney General, sets up the CalCompute consortium and has the Department of Technology recommend updates annually.

Safety topics

general-purpose / frontier AItransparency & labellingincident reportingcatastrophic riskaccountability & liability

Status notes

Signed 29 Sep 2025. Large frontier developers must send the Office of Emergency Services a summary of any assessment of catastrophic risk from internal use of their frontier models every three months or on another reasonable schedule they set (Business and Professions Code section 22757.12(d)); the Office runs a portal for these summaries and for critical safety incident reports. CalCompute is a consortium within the Government Operations Agency to develop a framework for a public computing cluster. The Governor's signing message says the Act provides a compliance pathway for the incident-reporting requirements if federal standards are adopted. Governor's EO N-9-26 (18 Sep 2026) seeks recommendations by 16 Nov 2026 on strengthening state law: independent verification of filings, onsite auditors, a 'kill switch', and loss-of-control incidents. The federal FRONTIER Act would preempt it.

Sources

Open the official source https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/

Secondary source: https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260SB53

Local copy

Archived. The secondary source could not be downloaded (HTTP 403 Forbidden). Last attempt 2026-10-16 16:38:54 UTC

Official source (third-party copy, not published) archive/US-CA-05/2026-10-16/official.html

Retrieved 2026-10-16 15:53:56 UTC

SHA-256 90850f8184638ba37f5d442570b59a63716b0385aea3be6cb2073396a6ec2d8a

Verification

The instrument's own official text, or an official register, was stored and confirms the title, date and status. checked 2026-10-16 · official-url-fetched

Stored: the Governor's release of 29 Sep 2025 (gov.ca.gov), which gives the title, the signing and the main provisions; the Governor's signing message of 29 Sep 2025 (a scan, read by OCR); and the Office of Emergency Services page for the Act's reporting portal (caloes.ca.gov), which quotes Business and Professions Code section 22757.12(d). leginfo.legislature.ca.gov refused automated downloads (HTTP 403, also in a browser), so the chaptered text, the chapter number and the operative date could not be stored. The effective date of 1 Jan 2026, the 10^26-operation compute threshold, the $500 million revenue threshold for large frontier developers, the duty to publish transparency reports for new models, the penalty cap of $1 million per violation and the incident-reporting deadlines (15 days, or 24 hours where there is imminent risk of death or serious injury) come from secondary sources and from the Assembly Privacy and Consumer Protection Committee's analysis of the bill as amended on 5 Sep 2025, not from the statute. Earlier notes cited a Fortune report of 14 Sep 2026 alleging that recent OpenAI releases may not comply; it rests on a press source only.

Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-16: 5 field(s) supported.

History

2026-10-18related: [{"slug": "ca-sb-1047", "id": "US-CA-03"}, {"slug": "ca-frontier-ai-policy-repor → [{"slug": "ca-sb-1047", "id": "US-CA-03"}, {"slug": "ca-frontier-ai-policy-repor — two-way link added: US-CA-01, US-CA-02, US-CA-06, US-CT-01 link(s) to US-CA-052026-10-18-fix-related-links · Claude Code (claude-opus-5-5)
2026-10-16notes: Signed 29 Sep 2025; effective 1 Jan 2026. Applies to frontier models trained abo → Signed 29 Sep 2025. Large frontier developers must send the Office of Emergency — research remarks removed ('secondary source', 'Not amended as of verification', 'commentators named it'), the unsupported thresholds, penalty cap and incident deadlines moved to verification.note, and the quarterly catastrophic-risk summaries and the signing message's compliance pathway added from stored texts2026-10-16-refresh-ids-us-04-us-ca-05-us-tx-01-ca-0 · Claude Code (claude-opus-5-5)
2026-10-16summary: First US state frontier-AI safety law. Large frontier developers must publish a → California's frontier AI safety law. Large frontier developers must publish a fr — rewritten so that every claim rests on a stored official text (the Governor's release and the Office of Emergency Services page); the duty to publish transparency reports for new models is dropped from the summary because no stored official text states it, and is recorded in verification.note2026-10-16-refresh-ids-us-04-us-ca-05-us-tx-01-ca-0 · Claude Code (claude-opus-5-5)

Related records

California SB 1047 California (US) · Lapsed

California Frontier AI Policy Report California (US) · Published (non-binding, current)

California EO N-9-26 California (US) · In force

California SB 813 / AB 1405 (AI independent verification & auditors) California (US) · Adopted, not yet in force

New York RAISE Act New York (US) · Adopted, not yet in force

EO 14365 (state AI law preemption order) United States · In force

FRONTIER Act United States · Draft / proposed

Illinois AI Safety Measures Act (SB 315) Illinois (US) · Adopted, not yet in force

California AB 2013 California (US) · In force

California AI Transparency Act California (US) · Partially in force

California SB 243 (companion chatbots) California (US) · In force

Connecticut Public Act 26-15 (SB 5, AI and online safety) Connecticut (US) · Partially in force

Citation

California State Legislature; signed by Governor Gavin Newsom. Transparency in Frontier Artificial Intelligence Act (SB 53) [California Senate Bill 53 (2025-2026 Regular Session)]. 2025-09-29. https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/. Status: In force; verified 2026-10-16 (level A). Local copy archive/US-CA-05/2026-10-16/official.html retrieved 2026-10-16 15:53:56 UTC, SHA-256 90850f8184638ba37f5d442570b59a63716b0385aea3be6cb2073396a6ec2d8a. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record US-CA-05.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.