California just built a registry for AI auditors
Newsom signed SB 813 and AB 1405 on September 9, creating a state registry for independent AI auditors and a framework for third-party verification of AI systems.
California signed two AI audit bills into law on September 9, and the practical effect is that "we had it independently reviewed" is about to become a claim with a definition behind it. Governor Newsom signed SB 813 and AB 1405, creating a framework for independent verification organizations and a state registry for the auditors who staff them.
What actually happened
SB 813, authored by Senator Jerry McNerney (D-Pleasanton), establishes a framework for independent verification organizations that can assess AI systems and models for compliance with state law. AB 1405, from Assemblymember Rebecca Bauer-Kahan (D-Orinda), creates the state registry for AI auditors and sets standards for their independence, transparency, and integrity.
Together they answer a question SB 53 — the Transparency in Frontier Artificial Intelligence Act, passed in 2025 — left open: who is qualified to check, and what makes them credible enough to be believed. McNerney's framing in the announcement is blunt about the motive: California is moving because Washington is not.
Note what these bills do not do. They do not impose new testing obligations on the average business deploying AI. They build the supply side — the auditors, the standards, the registry. The obligations that route work to those auditors are the existing and future ones.
Why an AI auditor registry matters for your business
You will meet this law through your vendors, not your own compliance team.
Right now, when a software vendor tells you their model is "independently evaluated," that phrase carries roughly the weight of "artisanal." There is no registry to check the evaluator against, no independence standard they had to clear, no way to tell a real audit from a logo on a PDF. AB 1405 puts a list somewhere. Lists are checkable.
Three things worth doing before the vendor conversation gets there first:
Ask who did the evaluation, by name. Not "a third party." The organization. Write it into the security questionnaire you already send.
Separate the claim from the scope. An audit of a model is not an audit of the product you are buying. Ask what was assessed and against what.
Keep the evidence yourself. If a vendor's compliance story ever gets tested, you want the attestation in your own storage, dated, not a link to a page they control and can quietly edit.
None of this is new discipline. It is the vendor diligence you already do for SOC 2, applied to a claim that until now had no floor under it.
Key takeaways
- Newsom signed SB 813 and AB 1405 on September 9, 2026
- SB 813 creates a framework for independent verification organizations that assess AI systems for compliance with state law
- AB 1405 creates a state registry of AI auditors with standards for independence, transparency, and integrity
- The bills build audit supply and infrastructure — they do not themselves impose new testing duties on most deployers
- Expect to encounter this through vendor claims: ask who evaluated the system, and what scope it covered
"Independently evaluated" is a sentence, not evidence. We build AI features with logged decisions, traceable inputs, and the kind of documentation that survives a diligence request. See how we ship AI you can actually account for, or bring us the vendor claim you cannot verify.
Sources: Office of Governor Gavin Newsom, AB 1405 bill record, CalMatters Digital Democracy.
- #ai-regulation
- #california
- #ai-audit
- #compliance
- #vendor-risk
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Suno v6 retrained on licensed music. Check your lineage
Suno rebuilt v6 from scratch on licensed Warner, BMG and Believe catalogs and retired its old models the same day. Training-data provenance is now a product feature.
Read itOpenAI tests outcome-based pricing after an 80% price cut
OpenAI's CFO says an 80% cut to Luna drove 10x usage, and the company is testing outcome-based pricing. Re-run your AI cost math before the meter changes.
Read it