California SB 53: requirements, deadlines and compliance actions
California SB 53: Transparency in Frontier Artificial Intelligence Act
What is the California SB 53?
SB 53 requires "large frontier developers" (developers of the most compute-intensive models above statutory thresholds) to publish a frontier AI framework describing how they assess and mitigate catastrophic risks, publish transparency reports when deploying new frontier models, report critical safety incidents to the California Office of Emergency Services, and protect employees who report safety concerns. It also directs creation of a public computing cluster ("CalCompute").
Status note: Signed 29 September 2025; operative 1 January 2026 for most provisions. A reviewer must confirm the operative dates and any federal pre-emption challenge.
Who does it apply to?
Applies to developers of frontier models meeting compute-based thresholds, with heightened duties for large frontier developers above a revenue threshold. Most obligations are transparency and reporting duties rather than prescriptive technical requirements.
Frontier-model developers operating in California, including companies based elsewhere that train frontier models and do business in the state.
- Sectors
- Cross-sector / all sectors
- AI use cases
- Generative AI and foundation models
- Risk categories
- Systemic-risk general-purpose model
- AI system types
- General-purpose AI modelGenerative AI system
When do the requirements apply?
Signed 29 September 2025; operative 1 January 2026 (to be confirmed).
| Date | Milestone | Source reference | Status |
|---|---|---|---|
| Operative date Most provisions operative; confirm on the official bill page. |
— | Passed confidence: medium |
What must organisations do?
Large frontier developers should publish a frontier AI framework and model transparency reports, set up critical safety incident reporting, and implement whistleblower protections and internal reporting channels.
Legal requirement Large frontier developers must publish a frontier AI framework Business and Professions Code, Chapter 25.1 (as added by SB 53)
Large frontier developers must publish and maintain a framework describing how they incorporate national and international standards, assess catastrophic risk, apply mitigations, secure model weights, and govern internal processes, and must review it at least annually.
Practical action: Publish a safety framework and keep a change log of revisions.
Evidence examples: Published frontier AI framework
Framework mapping (original, editorial): NIST AI RMF 1.0 GOVERN 1.x; NIST AI 600-1
Obligation pageApplies from 1 Jan 2026 Source-linked
Legal requirement Report critical safety incidents to the Office of Emergency Services Business and Professions Code, Chapter 25.1 (as added by SB 53)
Frontier developers must report critical safety incidents to the California Office of Emergency Services within the statutory time limit after discovery, and the Office is to establish a reporting mechanism.
Practical action: Define critical-safety-incident criteria and an escalation path with the statutory clock.
Evidence examples: Incident classification and reporting procedure
Framework mapping (original, editorial): NIST AI RMF 1.0 MANAGE 4.3
Obligation pageApplies from 1 Jan 2026 Source-linked
Penalties
Civil penalties enforced by the Attorney General for non-compliance, with amounts set in the statute (reviewer to confirm figures).
Official sources
-
SB 53 Transparency in Frontier Artificial Intelligence Act (bill text and history)
California Legislative Information · 29 Sep 2025 · Tier 1 source
Change history
- — California SB 53 frontier-model transparency duties become operative
- — California enacts the Transparency in Frontier Artificial Intelligence Act (SB 53)
Record version 1: Initial structured record; penalty amounts and chapter number to be confirmed.. Full edit history is in the GitHub repository.
Informational only, not legal advice. Verify every claim against the linked official sources and consult qualified counsel before acting.