Newsom Signs No Robo Bosses Act and 20+ AI Laws in California's 2026 Session
On Wednesday 30 September 2026, his last day to act on bills, California Governor Gavin Newsom signed a package of AI laws aimed at workplaces, health care and online content, including SB 947, the "No Robo Bosses Act", which stops employers relying solely on an automated system to discipline or fire a worker (AP via KSAT). KQED reports SB 947 takes effect on 1 July 2027. Across the session, a law firm tracking 25 privacy and AI bills counted 22 signed and three vetoed. The package was signed while the White House is still trying to get state AI rules preempted.
What happened
The workplace bills were the headline. According to KQED, Newsom signed five on 30 September:
- SB 947 (Sen. Jerry McNerney): a human must corroborate any discipline or firing decision made with an automated decision system, and the worker must be told. Gig drivers are not covered, because Proposition 22 classifies them as contractors.
- SB 951 (Sen. Reyes): employers must give notice when AI contributes to a mass layoff, relocation or termination, and the state Employment Development Department must study AI's effect on jobs.
- AB 1883: restricts workplace surveillance tools that infer emotional state or collect neural data.
- AB 1331: bans workplace surveillance in bathrooms.
- AB 1979: bars AI from replacing a licensed health professional's clinical judgment, and extends medical confidentiality to health chatbots.
Earlier, he had signed SB 1050 (disclosure when ads use an AI-generated performer) and AB 1609, which requires large businesses' customer service chatbots to make a good-faith effort to connect a customer to a human within 15 minutes, per KQED and Inside Global Tech. SFist reports he also signed, in September, a law setting up a system for independent groups to evaluate AI model safety, with a state registry and auditor standards; the Stauss list identifies these as SB 813 and AB 1405.
He vetoed AB 2575, which would have protected health workers who override AI recommendations they consider unsafe, and AB 2656, which would have required notice to public employees when AI performs work within their job scope (KQED). McNerney's summary of his own bill: "Nobody wants to be fired by a machine."
The numbers
| Bill | What it does | Effective / status | Source |
|---|---|---|---|
| SB 947 | Human corroboration and worker notice for AI-driven discipline or firing | Signed 30 Sep; effective 1 Jul 2027 | KQED |
| SB 951 | Notice when AI contributes to mass layoffs | Signed 30 Sep | KQED |
| AB 1609 | Customer service chatbots: no posing as human, route to a person within 15 minutes | Signed; applies to businesses over $500M revenue | Stauss Firm |
| SB 1000 | AI Transparency Act: removes the 1,000,000 monthly user threshold for covered providers | Signed | Inside Global Tech |
| SB 1119 | Companion chatbot protections for minors, independent child safety audits | Signed | Stauss Firm |
| SB 813 / AB 1405 | Independent verification organisations; AI auditor registry | Signed in September | Stauss Firm |
| AB 2575 | Protect health workers who override AI | Vetoed | KQED |
One count is worth flagging. Inside Global Tech says Newsom signed over 20 AI-related bills in the final days of September; the Stauss tally of 22 signed covers privacy and AI bills across the whole session. Law firm trackers differ on which bills they include.
Why it matters
The timing is pointed. AP reports that the day before, Trump said top tech leaders had signed a voluntary accord to "self-police" AI development. Newsom's response, also via AP, was that California has to "do a lot more in the absence of federal leadership." He also signed an executive order telling state agencies to keep saying "artificial intelligence" rather than "super intelligence", the term Trump had ordered US diplomats to use.
The substance matters more than the politics. Most earlier state AI laws focused on model developers or on consumer-facing disclosures. This batch regulates deployers, specifically employers and businesses running chatbots, which is where most companies actually sit. SB 1000 also widens the AI Transparency Act by dropping the user threshold, so smaller generative AI providers fall inside it.
And it is durable. Preemption has not passed Congress, and after the November midterms gave Democrats both chambers it looks less likely (I wrote that up in a separate article).
What sceptics say
Labor got part of what it asked for. Lorena Gonzalez of the California Federation of Labor Unions praised Newsom for starting to put "guardrails on AI at work", but two workplace AI bills were vetoed, and SB 947 is a narrower version of a bill Newsom vetoed a year earlier. Rideshare drivers, arguably the workers most managed by algorithms, are outside it entirely.
From the other side, a growing state patchwork is exactly what the December 2025 federal executive order on state AI laws set out to stop in the name of a uniform national standard. California employers will now have AI notice duties at the state level that do not exist in neighbouring states. And "a human must corroborate" can become a rubber stamp if the reviewer only sees the system's recommendation.
What it means for engineering and DevOps teams
My view: if you build HR, workforce management or customer service software that is used in California, these are product requirements now, with a deadline.
- Make the human step real and logged. SB 947 compliance will be argued from records: what the system recommended, who reviewed it, what else they looked at, when. Design that audit trail now, with the same discipline you would apply to audit logs for enterprise SaaS.
- Build the escape hatch for chatbots. AB 1609's 15-minute handoff is an availability and routing requirement. It needs queue metrics and alerting, not a disclaimer.
- Keep provenance metadata intact. SB 1000 and AB 2713 strengthen latent disclosure and provenance duties. Image pipelines that strip metadata on resize will need fixing.
FAQ
What is the No Robo Bosses Act?
It is California SB 947, signed on 30 September 2026. It bars employers from relying solely on an automated decision system to discipline or fire a worker, requires a human to corroborate the decision, and requires notice to the worker. It takes effect on 1 July 2027.
When do California's 2026 AI laws take effect?
Most California laws take effect on 1 January of the following year unless the bill says otherwise. SB 947 has a later date, 1 July 2027, according to KQED. Check each bill's text for its own operative date.
Did Newsom veto any AI bills in 2026?
Yes. KQED reports he vetoed AB 2575, on health workers who override AI recommendations, and AB 2656, on notifying public employees when AI does their work.
Related: Enterprise-Ready SaaS in 2026: SSO, SCIM, Audit Logs and the SSO Tax, Policy as Code for AI Agents in 2026 and AI Agent Failures in Production: 5 Incidents, 14 Failure Modes (2026).
Spot an error or something out of date? Email me and I will fix it and log it on the corrections page.
I'm Ajay Kumar — I build and operate PandaStack, an open-source Firecracker microVM cloud for AI agents, and write about the infrastructure and tooling around it.
Need this kind of infrastructure work? See what I do or email hello@ajayk.sh.
Related
Anthropic Lets You Self-Host Claude Managed Agents, in Private Beta
On 12 November 2026 Anthropic opened a private beta of self-hosted Claude Managed Agents: the API server, agent loop, session history and workers run on your infrastructure, and only model calls leave. What changes for platform teams.
6 minNov 10, 2026ARTEX: China-Linked Hackers Aimed an AI Pentest Framework at 2,250+ Tenants
Microsoft tied Storm-0940 to the ConnectWise supply chain compromise using the ARTEX agentic framework, and CrowdStrike says VAPOR PANDA used it on 200+ organisations in 11 days. What happened and what to check.
6 minNov 10, 2026Australia Plans AI Safety Regulator, 6 Mandatory Rules After OpenAI Breach
On 10 November 2026 Anthony Albanese announced an AI safety regulator, six mandatory guardrails and acceptance of all 18 review recommendations after an OpenAI agent got into a Medicare statistics portal in June and was disclosed in September.
6 min