California's No Robo Bosses Act Keeps a Person Accountable for Firing Decisions
Three new California laws signed September 30, 2026 bar employers from firing or disciplining workers on an automated system's output alone and ban emotion-tracking AI and restroom surveillance on the job.
Key Takeways
- California's No Robo Bosses Act (SB 947) bars employers from firing or disciplining a worker based solely on an automated system's output and requires a human to independently review the recommendation first, starting July 1, 2027.
- AB 1883, in force since September 30, 2026, bans workplace AI tools that recognize or predict employees' emotional states or collect neural data, with civil penalties of up to $500 per violation enforceable under PAGA.
- AB 1331 bans workplace surveillance tools from monitoring restrooms, with narrow exceptions for badge and alarm access systems that use no audio, video, or embedded AI.
- Under SB 947, workers must receive written notice when an automated system was used in an employment decision about them, and the California Labor Commissioner enforces the law.
Three new California laws now restrict AI-driven workplace surveillance and automated firing. Gov. Gavin Newsom signed them on September 30, 2026. The centerpiece is SB 947, the No Robo Bosses Act, which bars employers from firing or disciplining a worker based solely on an automated system's output. Two companion bills, AB 1883 and AB 1331, took effect the day they were signed. One bans workplace AI that reads employees' emotions or collects their neural data. The other keeps surveillance tools out of restrooms.
Together, the three bills answer a question every employer using these tools will face. When software recommends ending someone's job, who answers for the decision? From July 1, 2027, California's answer is a person.
What the three bills require
SB 947 was introduced by state Sen. Jerry McNerney of Pleasanton and sponsored by the California Federation of Labor Unions, AFL-CIO. It requires a human to independently review and verify an automated system's recommendation before any termination or discipline. It also bars employers from using these systems to infer protected characteristics or to predict and preempt retaliation claims. When an automated system is used in an employment decision, the worker receives written notice. The California Labor Commissioner enforces the law, as can the state Attorney General and local prosecutors.
| Bill | What it bans or requires | Effective date | Enforcement |
|---|---|---|---|
| SB 947, the No Robo Bosses Act | Requires human review and verification of an automated recommendation before any termination or discipline. Bars inferring protected characteristics or predicting retaliation claims. Requires written notice to workers. | July 1, 2027 | California Labor Commissioner, state Attorney General, or local prosecutors |
| AB 1883 | Bans workplace AI that recognizes or predicts emotional states or collects neural data. | September 30, 2026 | Up to $500 per violation, under PAGA |
| AB 1331 | Bans surveillance tools in restrooms. Narrow exceptions for badge and alarm systems with no audio, video, or embedded AI. | September 30, 2026 | Up to $500 per violation, under PAGA |
Rules written for a slower world, updated ahead of the practice
Most employment law was written for a slower world. A supervisor wrote the warning and signed the termination letter. Software now handles parts of that work and updates every few weeks. Tools that score productivity and recommend discipline are on the market, and approving their output takes one click.
What stands out about SB 947 is its timing. Sen. McNerney and the California Legislature wrote the human-review rule while employers are still deciding how much authority to give these systems. They also gave employers until July 1, 2027 to build the review step into their process. I would point any executive to that sequencing first. Governance that arrives ahead of a practice shapes how the practice grows. Governance that arrives later has to unwind habits people depend on.
I made a related case about AI agents inside large companies, where the same accountability question applies. When an agent acts, who answers for it?
Sen. McNerney put it plainly:
"No worker should ever be fired or disciplined by a robo boss. AI must remain a tool controlled by humans, not the other way around."
A person above the algorithm
Think about the supervisor who opens a dashboard after a long shift and finds a recommendation to discipline someone she has worked beside for years. Under the No Robo Bosses Act, she has to read the file herself and check the numbers against what she saw on the floor. The worker gets written notice that software played a part. Both of them now know where the decision came from, and the supervisor knows it is hers.
In my forthcoming book, The Great Relearning, I call this staying Above the Algorithm. Judgment and accountability stay with a person, even as the analysis moves to machines. SB 947 turns that idea into California AI employment law. Software can recommend. A person decides.
That shift asks something new of managers. Reading a model's recommendation critically, and knowing when to overrule it, is a skill many will learn as they go. Employers have about 9 months to teach it.
Where the companion bills draw firm lines
AB 1883 and AB 1331 use outright bans in place of a review requirement. AB 1883 covers tools that recognize or predict emotional states. It also covers "neural data," information measuring nervous-system activity. AB 1331 keeps monitoring out of restrooms, apart from narrow badge and alarm exceptions. Both carry civil penalties of up to $500 per violation under California's Private Attorneys General Act (PAGA), and both are already in force.
Connecticut chose a different mechanism. Its CART Act gives workers whistleblower rights when they report AI-related harm. One state limits what the systems may do to workers. The other protects the workers who speak up about them. For how these rules fit together, see the AI governance content hub.
Two questions to take into your next meeting
For employers, the months before the No Robo Bosses Act takes effect on July 1, 2027 are a chance to design the review step on your own terms. I bring two questions into my own conversations, including my advisory work on AI governance with the California State University system. You are welcome to borrow them. Which decisions about our people already start with a system's recommendation, and who reviews them today? If a worker asked us tomorrow how a decision was made, could we explain it in writing?
If you work in California and manage no one, the notice requirement answers a question of your own. Did software play a part in a decision about your job? After July 1, 2027, your employer has to tell you in writing.
Every one of these decisions reaches a kitchen table. A disciplinary note that started as a model's error can follow a person into the next review and the next job search. California has now said that a person stays accountable for the decisions that shape a working life. Which of your company's decisions would pass that test today? I'm easy to find.
Can my employer use AI to fire me in California?
Starting July 1, 2027, an employer can use an automated system as one input, but a human has to independently review and verify its recommendation before any firing or discipline, and you must receive written notice that the system was used. If you believe an employer skipped that review, the California Labor Commissioner is the office that enforces SB 947, alongside the state Attorney General and local prosecutors.
Is emotion-detection AI already banned in California workplaces?
Yes. AB 1883 took effect when it was signed on September 30, 2026, so there is no phase-in period for workplace AI tools that recognize or predict employees' emotional states or collect neural data. Employers that bought software suites before the law passed have reason to check whether any built-in feature falls into that category, since penalties of up to $500 per violation apply now.
Here is what makes Alex a credible voice on this topic: As an AI governance advisor to the California State University system, Alex works on the question SB 947 now answers in employment law, which is who stays accountable when software recommends a decision about a person.
Talk with Alex about AI accountability for your organization →
.svg)
