California Bans AI-Only Firing Decisions Under New No Robo Bosses Act

California just made it illegal to let an algorithm fire you without a human signing off. Governor Gavin Newsom signed SB 947 — the No Robo Bosses Act — on October 4, 2026, making California the first state in the country to require human oversight for AI-assisted termination and discipline decisions. For millions of workers whose performance is tracked by algorithmic systems, the law is the most direct employment protection from AI automation enacted anywhere in the United States.

Background

Workplace AI systems — software that monitors productivity, flags attendance patterns, scores customer service calls, and grades warehouse pick rates — have quietly become the de facto managers for a growing share of the American workforce. Amazon, Uber, Lyft, and countless logistics operators have faced lawsuits and legislative scrutiny over automated deactivations that left workers with no human to appeal to and no clear explanation of what triggered the decision. California has been the most active state on AI employment rules, passing data privacy protections and algorithmic accountability bills in recent sessions, but SB 947 goes further than any prior law by making the human-in-the-loop requirement a hard legal standard.

What the Law Does

SB 947, authored by State Senator Josh Becker (SD-13) and carried by Senator Jerry McNerney, imposes three core obligations on California employers:

• Human review required. No employer may rely solely on an automated decision system to terminate, suspend, demote, or formally discipline a worker. A human decision-maker must review the AI recommendation before any adverse action takes effect.
• Written notice to the worker. Employees subject to an AI-assisted disciplinary decision must receive a written notice identifying which data the system used, what factors it weighted, and which human reviewer approved the outcome.
• Right to explanation. Workers have the right to ask for — and receive — a plain-language explanation of why the AI flagged them.

The law applies to employers with 15 or more employees operating in California. It takes effect July 1, 2027, giving businesses roughly nine months to audit their HR tech stacks and update vendor contracts.

Penalties are civil: the California Labor Commissioner can levy fines of up to $1,000 per violation per employee, and workers retain a private right of action to sue for reinstatement and back pay.

Who It Affects

California employs roughly 19 million workers. The industries most directly affected are those where algorithmic management is already standard:

• Gig and app-based platforms — rideshare and delivery companies that deactivate drivers based on rating thresholds or completion rates
• Warehousing and logistics — fulfillment centers that score workers on pick-per-hour and flag outliers automatically
• Call centers and customer service — operations that use speech analytics to score interactions and generate disciplinary records
• Healthcare staffing — shift management platforms that flag attendance patterns and generate write-ups

Employers who currently use third-party HR software will need to verify that their vendors can produce the required data trail for each adverse action. That likely means contract amendments and, for some platforms, software updates.

What Employers Need to Do Before July 1, 2027

The nine-month runway is meaningful but not generous. Companies should move on three tracks:

1. Inventory AI touch points in HR workflows — identify every system that influences hiring, performance scoring, scheduling, or termination recommendations.
2. Designate human reviewers — document who is authorized to make the final call on any AI-flagged adverse action, and build that approval step into the workflow.
3. Update vendor agreements — require SaaS HR vendors to provide the data fields needed for written employee notices (inputs, weights, outputs, reviewer name, date).

Employment lawyers contacted by NexChron noted that the private right of action is the most potent enforcement mechanism. Unlike some state AI laws that rely entirely on agency enforcement, SB 947 gives individual workers standing to sue, which makes the statute self-enforcing in a state with a large, active plaintiff's employment bar.

Impact on HR Tech

The law lands at an awkward moment for the HR software industry. Vendors have spent the last three years selling AI hiring and performance tools as cost-reducing automation — the pitch being that software removes human bias and speeds decisions. SB 947 reframes that pitch: automation is a drafting tool, not a decision-maker.

Expect vendors to add "human review" workflow layers to California-market products, similar to the GDPR-driven consent toggles that European data laws forced on marketing platforms. Companies that sell purely automated adverse-action workflows will face the starker choice of either rebuilding the product or exiting California enterprise sales.

What to Watch

The immediate question is whether other large states follow California's lead. New York, Illinois, and Washington have all had similar bills in committee; California's enactment gives those efforts a working template and a signing precedent. At the federal level, the EEOC has signaled interest in AI employment guidance but has not proposed binding rules — SB 947 effectively sets a floor that federal action would need to meet or exceed to matter.

Watch also for the first enforcement actions after July 1, 2027. The written-notice requirement will generate a paper trail that plaintiff attorneys can use to reconstruct how an AI system reached a decision, which could expose liability beyond the termination itself if the underlying model is found to have discriminatory disparate impact.

Hector Herrera covers AI policy and workforce technology for NexChron.