Illustration: AI-generated conceptual image; it depicts neither the bill signing nor an actual worker incident.

California’s new workplace laws address three different uses of AI and automation. Rules on covered layoff notices and certain surveillance tools are expected to take effect January 1, 2027; the law governing automated disciplinary and termination decisions becomes operative July 1, 2027.

California Governor Gavin Newsom announced the signing of several workplace bills on September 30. His summary groups them as protections from AI and surveillance. The enrolled texts set out three distinct rules: one regulates particular decisions made with an automated decision system, another adds disclosures to existing large-layoff notices, and a third restricts specific AI surveillance uses.

The big change

  • What changed: California now sets separate checks on automated discipline and termination, adds technology details to already-required large-layoff notices, and limits AI emotion inference and neural-data collection at work. Each rule has its own trigger. (SB 947, SB 951, AB 1883)
  • Why it matters: An employee affected by a decision primarily based on an automated system can learn what data of theirs it used and reach a person about the decision. The Employment Development Department’s public summaries will show automation identified in covered displacement notices. Existing Cal/WARN thresholds leave other job losses outside that record.
  • What to watch: SB 951 and AB 1883 are expected to take effect January 1, 2027; SB 947 operates July 1. The department must publish notice summaries and report on AI and hiring by January 1, 2028. For any workplace decision, notice or tool, the relevant statutory trigger determines whether these new rules apply.

Automated discipline and termination require a corroborating person

SB 947 defines an automated decision system broadly enough to include computational processes based on machine learning, statistical modeling, data analytics or AI that produce a score, classification or recommendation used in a decision that materially affects people. It excludes tools such as spam filters, antivirus software, calculators and databases. Its decision rule concerns discipline and termination, rather than every use of AI in hiring or management.

Under Labor Code section 1522, an employer cannot rely solely on such a system for either decision. If it primarily relies on the system’s output, a human must corroborate the decision using the data behind that output or other relevant supporting information. The law names supervisor evaluations, personnel files, work product, peer reviews and witness interviews as possible sources. If the output cannot be corroborated, or the reviewer finds it inaccurate, incomplete or misleading, the employer cannot use it for that decision.

In that primary-reliance situation, the employer must give the affected employee a separate written notice when communicating the decision. It must say that a person reviewed and corroborated the output, provide a human contact, and explain the employee’s right to request a meaningful, objective description of their own data used by the system. Descriptions must anonymize other people’s personal information. Section 1522 also forbids using these systems to infer an employee’s protected status, predict and penalize a worker for exercising legal rights, or prevent compliance with labor and civil-rights laws.

The statute covers employers as defined in the bill, including public bodies, subject to its terms. It excludes parties to a collective bargaining agreement only when the agreement explicitly waives the law, sets out work terms and protects against algorithmic management. A separate exception applies to automated-system use to the extent required by, or reasonably necessary to comply with, a federal statute, federal regulation or binding federal contract relating to the development of aircraft for use in the national airspace or the development of products or services for national-security, military, space or defense purposes. It applies only to operations covered by that federal requirement.

Automation details attach to existing Cal/WARN notices

SB 951 amends California’s existing 60-day notice rule for a mass layoff, relocation or termination at a covered establishment. A covered establishment employs, or employed within the previous 12 months, at least 75 people. A mass layoff means 50 or more employees in any 30-day period at such an establishment. The statute separately defines relocation as moving all or substantially all operations at least 100 miles and termination as cessation or substantial cessation of operations. Its definition of an employee for notice purposes requires at least six months’ employment in the previous 12.

When an employer already owes notice under Cal/WARN and the event is caused in whole or substantial part by AI or other automated technology replacing or automating positions, the notice must identify the number, job classifications or occupations and work locations of affected layoffs, the functions to be automated, and the type of technology involved. It must carry the statement “This notice is for a technology displacement” at the top. The Employment Development Department must publish summaries of such notices and a quarterly statewide summary.

The amendment does not make every AI-related job loss a notice event. It also directs the department to report to the Legislature by January 1, 2028 on AI and hiring practices, including possible recommendations for a future notice requirement when employers permanently stop hiring or contracting for a position because of automation. That prospective notice requirement is a subject for the report, not a duty created by this bill.

The existing chapter also excludes specified temporary project and seasonal employment, and waives notice when an event is necessitated by a physical calamity or act of war. Those limits continue to matter when technology is involved.

AI emotion inference and neural-data collection face limits

AB 1883 applies to a workplace surveillance tool that uses AI on employees. Labor Code section 1581 prohibits using it to recognize, infer or predict a person’s emotional state, or to collect neural data. The bill defines neural data as information generated by measuring activity in an employee’s central or peripheral nervous system, excluding an inference from non-neural information. Its definition of employer includes governmental entities and labor contractors.

The text says employers may use workplace surveillance tools to ensure safety, or for other uses that do not meet the two prohibited specifications. It also has an exception for the development of aircraft for use in the national airspace or the development of products or services for national-security, military, space or defense purposes. That exception applies only to operations covered by a federal statute, regulation or binding federal contract where use of the tool is reasonably necessary to comply with that federal requirement. The safety clause is stated separately from the restriction on other uses; the bill does not prohibit all workplace monitoring.

The governor grouped AB 1331 with these measures. That separate bill addresses surveillance in workplace bathrooms, including tools without AI. Its bathroom rule is not an AI-specific ban.

The dates differ

The three bills are marked non-urgency in the Legislature’s status records. Article IV, section 8(c) of the California Constitution points to January 1, 2027 for SB 951 and AB 1883. SB 947 expressly says its new Labor Code part becomes operative July 1, 2027. The Legislature’s AB 1883 status page had not yet recorded the governor’s approval when checked October 1; the governor’s September 30 signing announcement supplies that confirmation.

Sources & further reading