California's Workplace AI Laws Raise the Bar for Employer Accountability

On September 30, 2026, California enacted three new laws regulating workplace AI: SB 947, AB 1883, and SB 951. Together, they prohibit employers from relying exclusively on AI to discipline or terminate employees, restrict certain forms of AI-powered employee monitoring, and require additional disclosures when AI or automation contributes to workforce reductions. These laws underscore a broader new reality: employers must fully understand where AI is already embedded in their employment systems and how it is used. AB 1883 and SB 951 take effect January 1, 2027; SB 947's requirements apply beginning July 1, 2027.

Employers operating in multiple states already face a growing patchwork of AI regulation. Illinois regulates certain uses of AI in employment decisions. New York City requires bias audits and notice requirements for covered automated employment decision tools. Colorado recently implemented broad AI requirements governing high-risk AI systems. California's latest legislation continues that trend and provides a preview of the questions regulators, agencies, plaintiffs' attorneys, and employees are increasingly asking: What AI systems are you using? What data are they analyzing? Who is overseeing the results? And can a human explain the decision that was ultimately made?

For many organizations, the greatest compliance risk is not intentional misuse of AI, but rather, it is discovering too late that existing technology already performs functions subject to regulation. Performance management platforms may generate disciplinary recommendations. Productivity tools may create employee risk scores. Monitoring software may include sentiment-analysis features that management never realized were activated. Workforce planning initiatives may eventually trigger disclosure obligations when automation replaces job functions.

The most effective response is not waiting for regulators to ask questions. It is understanding now where AI is already embedded within employment-related systems.

I. Human Judgment Must Remain Part of Discipline and Termination Decisions

California's new SB 947 prohibits an employer from relying solely on an automated decision system (ADS) to discipline or terminate an employee. If the employer primarily relies on an ADS output, a human must corroborate the decision using the underlying data or other relevant evidence. The employer cannot use an output that it cannot corroborate or that the reviewer concludes is inaccurate, incomplete, or misleading. For example, a manager reviewing an automated recommendation to suspend a worker should examine the records supporting that recommendation.

When an employer primarily relies on an ADS for discipline or termination, the affected employee must receive a separate, plain-language written notice when informed of the decision, in the language used for routine employee communications. It must be delivered through a simple, easy-to-use method. The notice must disclose primary reliance on an ADS, confirm human review and corroboration, provide contact information for a human who can explain the decision, explain the right to request a description of the employee's own data used by the system, and state that retaliation for exercising statutory rights is prohibited. A requested data description must be provided in a manner that keeps the individuals' personal information anonymized.

To ensure compliance, employers should pay particular attention to automated attendance and absenteeism systems, productivity scoring tools, call-center quality monitoring platforms, warehouse or logistics performance analytics, employee conduct detection software and AI-generated performance rankings. If an employer cannot explain the decision to terminate or discipline an employee without relying on an AI score or recommendation, then additional review procedures may be needed.  

II. Employers Must Evaluate Employee Monitoring Technologies

California's new AB 1883 prohibits employers from using AI-powered workplace surveillance tools on employees to recognize, infer, or predict emotional states, or to collect neural data. Surveillance tools include video and audio monitoring, continuous incremental time tracking, geolocation, and other devices that collect employee information by means other than direct observation by a person. Neural data means information generated by measuring activity in an employee's central or peripheral nervous system, excluding information inferred from non-neural data.

Given this law, employers should examine the actual capabilities of monitoring products. Features marketed as sentiment analysis, engagement measurement, or behavioral insights warrant review. A product's safety purpose does not, by itself, exempt prohibited emotion inference or neural-data collection.

III. Automation Projects May Now Create Additional WARN Considerations

SB 951 expands California Worker Adjustment and Retraining Notification Act (Cal/WARN). When an otherwise covered mass layoff, relocation, or termination is caused "in whole or in substantial part" by AI or other automated technology replacing or automating employment positions, the notice must additionally contain:

• The number, classification or occupation, and work location of layoffs substantially attributable to AI or automation.
• The job functions performed by the replaced workers that will be automated.
• The specific category or type of AI system or other automation that is substantially responsible for the displacement.
• The statement "This notice is for a technology displacement" at the top of the notice.

The new disclosures do not create a notice obligation for every automation project. For a covered workforce reduction, however, employers should assess and document technology's role early, especially when automation and other business reasons contribute to the same decision.

IV. Five Questions Every Employer Should Be Asking

Rather than treating these laws as three separate compliance projects, employers should view them as an opportunity to establish a broader workplace AI governance strategy.

Employers should be able to answer five questions:

1. What AI systems are currently being used across the organization?
2. Which employment decisions are influenced by those systems?
3. What employee data is being collected, analyzed, or scored?
4. Who is responsible for reviewing AI-generated recommendations?
5. How would we explain and defend those decisions if challenged by a regulator, employee, or jury?

If employers are unable to answer these questions, an AI governance review should likely be a near-term priority.

V. Looking Ahead

The larger lesson for employers is that workplace AI governance is quickly becoming an employment-law issue, not merely a technology issue. The organizations best positioned for the next wave of regulation will not necessarily be those using the least AI. They will be the organizations that know exactly where AI is being used, understand its capabilities, and have established clear governance processes around its use.

A good starting point for most employers is a Workplace AI Risk Assessment. Such an assessment can help identify AI-enabled systems already embedded in HR, productivity, security, workforce-management, and operational platforms; evaluate where those systems influence employment decisions; and develop practical governance measures before new regulatory requirements take effect.

For assistance evaluating workplace AI technologies, conducting an AI risk assessment, reviewing vendor relationships, updating employment decision procedures, or developing an enterprise AI governance framework, please contact Jennifer K. McCarty, Dean J. Shauger, Samuel C. Carey, or any member of Baker Donelson's Artificial Intelligence and Labor & Employment teams.