Gov. Gavin Newsom signed SB 813 and AB 1405 on Sept. 9, creating the first U.S. state framework for independent third-party AI audits and a registry of certified auditors. Read the fine print and the frame widens fast: this isn’t only a leash on frontier labs.

SB 813, carried by Sen. Jerry McNerney, stands up the California Artificial Intelligence Standards and Safety Commission inside the Government Operations Agency and empowers it to designate Independent Verification Organizations. AB 1405, from Assemblymember Rebecca Bauer-Kahan, sets up the auditor registry and the independence standards those verifiers must meet. The statutory hinge is SB 813’s definition of a “covered AI audit” as one assessing whether an AI system meets minimum safety, efficacy, reliability, or security requirements under state law. That language reaches downstream. Any AI system deployed in California that touches those criteria is potentially in scope, not just the models trained by the largest labs.

OpenAI and Anthropic backed the bills. That’s the tell. The frontier developers already sit inside SB 53, the 2025 transparency law, and industry support for a state-run auditor regime looks a lot like the 2002 Sarbanes-Oxley moment, when large incumbents welcomed the audit apparatus that raised costs for everyone else.

The practical read for a 5–30 person company using AI for lead scoring, applicant screening, or dynamic pricing: procurement questionnaires are about to grow a new line item. B2B and HR buyers will start asking whether the vendor’s system has been verified by a registry auditor.

More is coming. Newsom is still working through the 24 AI bills California sent him under a 30-day clock, and the Sept. 30 deadline on AI ad disclosure is still pending. Wednesday was the opening move.

Sources

Sources