A wave of new California legislation will reshape how employers—and the attorneys who represent them—may use artificial intelligence. On September 30, 2026, Governor Gavin Newsom signed four measures into law: SB 574, AB 1883, SB 951, and SB 947. Together, these legislative actions will regulate the use of artificial intelligence (“AI”) by employers in managing their workforces and by attorneys and arbitrators engaged in the practice of law within the state. Three of the four laws take effect January 1, 2027; the fourth, SB 947, follows on July 1, 2027.
For any company with operations or employees in California, these laws are not abstract policy statements. They impose concrete obligations—new notices, human-review requirements, recordkeeping, and disclosure duties—backed by civil penalties and public enforcement. Employers that already use AI to screen resumes, monitor productivity, flag performance issues, or plan restructurings should understand what each law requires well before the effective dates arrive.
AB 1883 – A Ban on AI Monitoring of Workers’ Emotions and Brains
AB 1883 takes direct aim at a new frontier of workplace surveillance. It prohibits employers from using AI-powered workplace surveillance tools for two purposes: (1) recognizing, inferring, or predicting an individual’s emotional state; and (2) collecting “neural data.” The law reaches these capabilities even when they are embedded in otherwise routine monitoring technology.
The statute defines its terms expansively. A “workplace surveillance tool” is one used to collect employee data, activities, communications, actions, behaviors, or biometrics through video or audio surveillance, continuous incremental time-tracking, geolocation, electromagnetic tracking, photoelectronic tracking, a photo-optical system, or other means. “Neural data” is defined as information produced by measuring activity in the central or peripheral nervous system.
The Labor Commissioner or a public prosecutor may enforce AB 1883, and violations carry a civil penalty of up to $500 per violation. The law contains two important exemptions. The first applies where the use of a workplace surveillance tool is reasonably necessary to comply with federal law and involves aviation, national security, military, or defense purposes. The second—and the one most employers will actually invoke—permits the use of a workplace surveillance tool to ensure safety. That safety exemption is particularly significant for law enforcement, transit, transportation, and other safety-sensitive employers that rely on fatigue- and impairment-detection technology; employers intending to lean on it should be prepared to document the specific safety risk addressed and confirm that collected data is not repurposed for other uses.
SB 951 – Disclosure of AI-Driven Mass Layoffs
SB 951 brings AI into California’s existing plant-closing framework. It amends the California WARN Act to require specific new disclosures when a mass layoff, relocation, or termination results from an AI system “in whole or in substantial part.” Like the balance of the WARN Act, the new requirement applies to employers that have employed 75 or more people within the preceding 12 months.
When AI drives the workforce reduction, the required notice must carry a conspicuous statement at the top of the page: “This notice is for a technology displacement.” The notice must also identify the number of employees affected, their occupations, and the layoff locations; the job functions performed by the displaced workers that will be automated; and a description of the AI system that resulted in the displacement. The practical effect is a new compliance trap—for employers that automate roles, because the “in whole or in substantial part” trigger is not defined and will require a careful, case-by-case judgment about causation before notices go out.
SB 947 – Human Decision-Making in Discipline and Discharge (the “No Robo Bosses Act”)
SB 947 is the most far-reaching of the four for day-to-day employment decisions, and it applies to both public and private employers without regard to the number of employees. It governs the use of “automated decision systems” (“ADS”) in discipline and discharge. An ADS is defined as any computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence that issues simplified output—a score, classification, or recommendation—used to assist or replace human discretionary decision-making and that materially impacts natural persons.
The centerpiece of the law is a human reviewer requirement. An employer may not rely solely on an ADS when making a disciplinary or termination decision. Where an employer primarily relies on ADS output to make such a decision, it must direct a human to corroborate the decision using the data collected or used to produce the output, or other relevant supporting information. If the output cannot be corroborated, or if the human reviewer determines it was inaccurate or deficient, the employer may not use it.
SB 947 also adds transparency obligations. When an ADS is used in a discipline or termination decision, the employer must, on the employee’s request, provide a written post-use notice describing the data the ADS primarily relied upon. The Labor Commissioner or a public prosecutor may enforce the law, with a civil penalty of up to $500 per violation. The requirements do not apply where the parties are covered by a valid collective bargaining agreement that contains an explicit waiver of the law’s provisions. SB 947 does not take effect until July 1, 2027, giving employers additional lead time to adapt.
SB 574 – Use of Generative AI by Attorneys and Arbitrators
SB 574 places guardrails on how licensed attorneys and arbitrators may use generative AI. At its core, the law prohibits attorneys from delegating the practice of law to generative AI. The statute defines “generative artificial intelligence” broadly, as a system that can generate derived synthetic content—including text, images, video, and audio—that emulates the structure and characteristics of the system’s training data.
The law layers several specific duties on top of that general prohibition. An attorney using generative AI must not enter personal identifying information or other nonpublic information into a system whose access is not restricted to the attorney and persons the attorney has authorized under obligations to protect the confidentiality of that information. In practice, that bars feeding client confidences into public, consumer-facing AI tools. Attorneys must also take reasonable steps to confirm the accuracy of AI-generated material and to address any “hallucinated output”—the fabricated cases and quotations that have already drawn judicial sanctions nationwide. An attorney may not file any document containing a citation the attorney has not personally verified, and the use of generative AI must be disclosed for documents submitted to the court.
The law reaches arbitrators as well. SB 574 specifically prohibits arbitrators from delegating any aspect of the decision-making process to generative AI, and bars them from relying on AI-generated information outside the record without first making reasonable disclosure to the parties.
What Employers Should Do Now
These laws share a common theme—human oversight, transparency, and restraint in how AI touches employees—but each demands its own compliance steps. Employers with a California presence should consider the following before the effective dates:
- Inventory current AI tools. Catalog every system that screens, scores, monitors, disciplines, or helps terminate workers, along with the vendors behind them. You cannot comply with laws you have not mapped to your actual technology stack.
- Isolate emotion- and neural-monitoring features. Review surveillance, wellness, and productivity tools for any capability that infers mood, stress, or sentiment, or that measures nervous-system activity, and disable those functions unless a statutory exemption clearly applies.
- Build a human-review protocol for discipline and discharge. Where ADS output informs disciplinary or termination decisions, establish a documented process for independent human corroboration—and a path to disregard output that cannot be verified.
- Update WARN procedures. Revise layoff-planning checklists to flag technology-driven reductions, capture the required “technology displacement” disclosures, and resolve the “in substantial part” causation question before notices issue.
- Revisit litigation and filing practices. Confirm that outside counsel and in-house attorneys verify every AI-assisted citation, protect confidential information from public AI tools, and disclose AI use in court filings as SB 574 requires.
- Coordinate with labor counsel on CBAs. Where a unionized workforce is involved, evaluate whether an express SB 947 waiver belongs in the next round of bargaining.
A New Compliance Landscape for AI at Work
California has again positioned itself at the leading edge of workplace regulation, and these four laws will not be the last word. As AI becomes further embedded in hiring, management, and separation decisions, employers should expect additional legislation, agency guidance, and litigation testing the boundaries of each statute. The employers best positioned for this new era are those that treat AI governance not as a one-time project but as an ongoing compliance function—documenting their tools, preserving human judgment in consequential decisions, and building transparency into the process before regulators or plaintiffs demand it.
The St. Louis employment attorneys at McMahon Berger have been representing employers across the country in labor and employment matters for over seventy years and are available to discuss these issues and others. As always, the foregoing is for informational purposes only and does not constitute legal advice regarding any particular situation as every situation must be evaluated on its own facts. The choice of a lawyer is an important decision and should not be based solely on advertisements.
