View source for News-California-SB-813-AI-Safety-Commission-2026

'''California SB 813''' would establish the '''California AI Standards and Safety Commission''' within the state Government Operations Agency to oversee AI safety regulation, including designating independent verification organizations and coordinating AI risk assessment across state agencies.<ref name="transparency">[https://www.transparencycoalition.ai/news/ai-legislative-update-april3-2026 Transparency Coalition, "AI Legislative Update — April 3, 2026"]</ref><ref name="digitaldemocracy">[https://calmatters.digitaldemocracy.org/bills/ca_202520260sb813 CalMatters Digital Democracy, "SB 813"]</ref>

== Bill Provisions ==

The bill adds Chapter 14, commencing with Section 8898, to Division 1 of Title 2 of the California Government Code. Key provisions include:<ref name="legiscan-text">[https://legiscan.com/CA/text/SB813/id/3299977 LegiScan, "SB 813 Bill Text"]</ref>

* '''Commission establishment''': Creates the California AI Standards and Safety Commission within the Government Operations Agency
* '''Liaison relationships''': Requires the Commission to maintain formal liaison relationships with state agencies deploying or procuring AI
* '''Technical expertise''': The Commission would provide AI technical expertise and risk assessments
* '''Independent Verification Organizations (IVOs)''': Authorizes the Commission to designate one or more IVOs, requiring them to implement AI risk mitigation plans covering cybersecurity, chemical/biological/radiological/nuclear threats, malign persuasion, and model autonomy/exfiltration
* '''Reporting''': IVOs must submit annual reports to the Legislature and Commission on evaluation resources, mitigation adequacy, and risks
* '''Fee authority''': The Commission may establish fees for applicants and IVOs to offset costs
* '''Multistakeholder Regulatory Organizations (MROs)''': The Attorney General would designate MROs for AI model and application certification, decertification, and reporting<ref name="legiscan-text"/><ref name="digitaldemocracy"/>

== Legislative History ==

* '''February 21, 2025''': Bill introduced by Senator McNerney<ref name="trackbill">[https://trackbill.com/bill/california-senate-bill-813-california-ai-standards-and-safety-commission-independent-verification-organizations/2672278/ TrackBill, "California SB 813"]</ref>
* '''May 12, 2025''': Placed on Senate Appropriations suspense file
* '''January 5, 2026''': Author's amendments read; re-referred to Senate Appropriations Committee<ref name="transparency"/>
* '''January 22, 2026''': Senate Appropriations approved (5-2, "Do pass")<ref name="trackbill"/>
* '''January 27, 2026''': Passed the Senate '''31-7'''; ordered to the Assembly<ref name="transparency"/><ref name="trackbill"/>
* '''As of April 13, 2026''': Held at Desk in the Assembly, pending referral<ref name="transparency"/>

== Current Status ==

SB 813 has passed the Senate and is pending in the Assembly, where it awaits committee referral. The Transparency Coalition's April 10, 2026 update noted the bill was "revived and re-referred to the Senate Appropriations Committee" after earlier passage, though the bill had actually already cleared Appropriations and the full Senate; it is now awaiting Assembly action.<ref name="transparency"/>

== Significance ==

If enacted, SB 813 would create one of the most comprehensive state-level AI regulatory bodies in the United States, with authority over AI safety verification, risk assessment, and coordination with other state agencies. The bill's IVO framework mirrors concepts from the California [[SB 1047]] debate but focuses on a commission-based oversight model rather than direct developer liability.

== References ==
<references />

[[Category:State Legislation]]
[[Category:California]]
[[Category:Risk-Based Framework]]
[[Category:Federal Regulation]]