In the last days of September 2026, Governor Gavin Newsom signed a package of California workplace bills that restrict how employers use AI. SB 947, the No Robo Bosses Act, requires a human to independently confirm a discipline or firing decision when the employer primarily relies on an AI tool, effective July 1, 2027. SB 951 adds technology-displacement disclosures to California WARN notices from January 1, 2027. AB 1883 bars AI workplace tools from inferring emotions or collecting neural data, also from January 1, 2027. HR teams with California employees should inventory tools now. The operative dates are in 2027, not this week.
California’s new workplace AI statutes, signed at the end of September 2026, do not copy Connecticut’s October 1 WARN disclosure. They add a human sign-off before AI-driven discipline, a specific technology-displacement warning on mass layoffs, and a ban on emotion and neural surveillance. HR, employee-relations, and people-analytics leads with any California headcount should map tools against those three duties before 2027 plan locks.
This briefing is operational context from the enrolled bills as described by employment counsel and contemporaneous reporting. It is not legal advice. Confirm effective text with California counsel before changing a policy.
What changed?
Newsom signed SB 947, SB 951, and AB 1883. Together they cover AI-based discipline and firing, WARN notices when automation drives a mass layoff, and AI tools that read emotions or neural data.
When?
Signed in late September 2026. SB 951 and AB 1883 take effect January 1, 2027. SB 947 takes effect July 1, 2027. Penalties described by counsel run to $500 per violation on the discipline and emotion measures.
Who is affected?
Employers with California employees using ranking, monitoring, contact-center, or people-analytics tools. SB 947 applies to current employees, not applicants.
What to do this week?
List tools that score, flag, or recommend discipline. Flag any sentiment or emotion feature aimed at workers. Do not wait for July 2027 to discover the feature is on by default.
What did the governor sign?
Bloomberg Law reported on September 30, 2026 that Newsom signed AB 1883, restricting employers from using AI-powered surveillance to collect neural data or infer emotional states, with civil penalties up to $500 per violation and enforcement by the labor commissioner and public prosecutors. The governor’s signing statement asked legislators to clarify the scope of AB 1883 and of AB 1331, a related restroom-surveillance measure.
Fisher Phillips, in an October 1 client note, placed AB 1883 alongside two other AI employment bills from the same signing window. SB 947, called the No Robo Bosses Act, requires a person to independently confirm the decision before an employer primarily relies on an AI tool to discipline or terminate a current employee. The reviewer must corroborate the tool’s output with supervisory evaluations, personnel files, work product, or similar evidence, and must give the employee a standalone written notice when the decision is delivered. Employees can request a description of their own data the system used over the prior 12 months. Counsel describes a $500 penalty per violation and a July 1, 2027 effective date. The law does not cover applicants.
SB 951 amends the California WARN Act. From January 1, 2027, a mass layoff, relocation, or termination caused in whole or in substantial part by AI or other automation needs four added disclosures: the statement “This notice is for a technology displacement” at the top, the number and type of affected positions, the job functions being automated, and the category of technology responsible. The state will publish a quarterly summary of technology displacements.
How do the three duties work in practice?
SB 947 is a process rule, not a ban on software. A manager can still look at an AI score. If that score is the primary basis for discipline or firing, a human has to confirm it with independent evidence and hand over a standalone notice. A dashboard used as one input among documented performance reviews is a different fact pattern from a tool that auto-generates the termination recommendation the company then adopts. Employee-relations teams should write that distinction into the 2027 playbook before the effective date, not during the first contested firing.
SB 951 is a notice rule tied to causation. “In whole or in substantial part” will be the phrase plaintiffs and unions read. HR should decide, with counsel, what evidence the company will keep when a restructure mixes automation with demand, site closure, or cost cuts. The required header is specific. A generic WARN letter will not satisfy it if the layoff meets the trigger.
AB 1883 reaches tools that recognize or predict an employee’s emotional state, and tools that collect neural data. Employment counsel warns that sentiment features in contact-center, collaboration, and HR analytics platforms can fall in, not only EEG headbands. There is a safety exception whose width is not yet clear. The bill dropped an express private right of action. Counsel still expects PAGA theories to be tested. Newsom asked the legislature to clarify scope, so the text employers train on in 2027 may not be the last word.
Why is this not the same as Connecticut’s October rule?
Connecticut’s first employer duty under its AI statute took effect October 1, 2026, and asks whether a WARN notice relates to AI or other technological change. California’s SB 951 is later, starts January 1, 2027, and prescribes particular words and fields, including a public quarterly summary. SB 947 has no Connecticut twin in that October rule. Multi-state employers should not paste one disclosure onto both notices.
Newsom vetoed a broader AI employment bill in 2025. The 2026 package is narrower and delayed. That delay is the planning window. It is not a reason to leave sentiment analysis switched on in a California contact center through next summer.
What should HR do before the effective dates?
Inventory systems that recommend discipline, rank performance, score calls for emotion, or monitor focus. Ask vendors, in writing, whether a California deployment can turn emotion or sentiment inference off. Record the answer in the vendor file. A safety exception is not a default setting.
Draft the SB 947 notice and the evidence checklist a reviewer must complete. Identify who is allowed to be the human confirmer. A confirmer who only clicks “agree” with the model is a weak file if the statute requires independent corroboration.
Update the California WARN template with the technology-displacement header and the four fields, and decide who declares that automation was a substantial cause. Legal should own that sentence. HR should own the headcount and job-function attachments.
Tell people analytics not to build 2027 scorecards that infer mood from email or chat for California employees until counsel has reviewed AB 1883’s final scope. The penalty figure cited by Bloomberg and Fisher Phillips is up to $500 per violation. Volume makes that more than a rounding error.
What should HR watch next?
Watch cleanup bills. Newsom’s signing statement asked legislators to clarify AB 1883 and AB 1331. A clarification in 2027 could narrow or widen the emotion ban. Watch the first quarterly technology-displacement summaries once SB 951 is live. They will show how peer employers describe automation. And watch PAGA complaints testing the emotion and surveillance statutes even without an express private right of action.
The dates to put on the HR calendar are January 1, 2027 for WARN disclosures and the emotion ban, and July 1, 2027 for the human sign-off on discipline and firing. The work to put on this month’s agenda is the tool inventory.
When did Newsom sign the workplace AI bills?
Bloomberg Law reported the AB 1883 signature on September 30, 2026. Fisher Phillips described the wider package, including SB 947 and SB 951, as signed over the days leading into its October 1 note.
When does the human sign-off rule start?
Fisher Phillips says SB 947 takes effect July 1, 2027. It applies when the employer primarily relies on an AI tool to discipline or terminate a current employee. Applicants are outside the statute as described.
What must a California WARN notice add?
From January 1, 2027, if AI or automation causes a covered layoff in whole or in substantial part, the notice must say “This notice is for a technology displacement” and include positions, functions being automated, and the technology category.
Are sentiment tools banned?
AB 1883 bars AI workplace surveillance used to recognize or predict an employee’s emotional state or to collect neural data, effective January 1, 2027, with a safety exception whose scope is still being debated. Counsel flags contact-center and HR analytics features as in range.
What are the penalties?
Bloomberg Law reported civil penalties up to $500 per violation for AB 1883. Fisher Phillips describes a $500 penalty per violation for SB 947 as well.
Does this replace Connecticut’s AI WARN disclosure?
No. Connecticut’s disclosure duty started October 1, 2026 and uses different wording. Multi-state employers need both templates.
Son Güncelleme / Last Updated: October 3, 2026
Related: Connecticut AI WARN disclosure · August JOLTS openings · HR hub
Discover more from Kurums | Business Intelligence
Subscribe to get the latest posts sent to your email.