California AI Employment Laws for 2027: What Business Owners Should Review Now

California employers face new AI and automated-decision rules in 2027. Learn what SB 947, SB 951 and workplace-surveillance changes mean for businesses.

Table of Contents

Your workplace uses AI already. Are your policies ready for 2027?

A company uses software to rank applicants, measure employee productivity or recommend disciplinary action. Managers see it as operational technology. California regulators increasingly treat these tools as part of the employer’s legal decision-making process.

California businesses should review their automated employment decisions before new 2027 requirements take effect. Existing employment-discrimination regulations concerning automated decision systems became effective October 1, 2025. Additional legislation signed in September 2026 addresses AI-assisted workplace decisions, surveillance and certain technology-related layoffs, with several new requirements arriving in 2027.

The details matter: not every automated spreadsheet is treated the same way as a system making disciplinary recommendations, and not every employer is covered by every statute. This article identifies the areas where a practical compliance review is warranted—not a universal prohibition on using AI.

The rules that were already in effect before 2027

California’s Civil Rights Council regulations governing automated decision systems in employment took effect October 1, 2025. The California Civil Rights Department explains that these regulations clarify how existing antidiscrimination law applies when employers use automated systems to make or inform employment decisions.

Among other matters, the regulations address discriminatory outcomes involving protected characteristics, potential unlawful medical inquiries and the retention of relevant employment records, including certain automated-decision data, for at least four years. They are a reminder that using a vendor does not automatically remove an employer’s responsibility for its own decisions.

Even before considering new legislation, an employer should know which tools affect recruiting, performance assessment, scheduling and termination, and what evidence is available to test their effects.

SB 947: restrictions on automated workplace decisions beginning July 1, 2027

California enacted Senate Bill 947 on September 30, 2026 (Chapter 859). The law’s relevant employment automated-decision provisions become operative July 1, 2027.

Under the enacted framework, employers face restrictions on using automated decision systems for prohibited purposes and on relying solely on those systems for certain disciplinary, termination or deactivation decisions. Where an employer primarily relies on an automated system for covered disciplinary or termination decisions, meaningful human review and specified notice obligations may apply. The law also addresses access to certain data about affected workers, anti-retaliation protections and statutory exceptions.

This is not a ban on using AI to assist a manager. It makes the structure of the decision important: what the system recommends, what the human actually reviewed, whether the review is genuine, and what explanation or notice the employee receives when the statute requires it.

What counts as an automated decision system?

SB 947 provides a statutory definition focused on computational processes derived from machine learning, statistical modeling, data analytics or AI that generate a simplified output—such as a score or recommendation—used to assist or replace human discretion and materially affect people. The final text also excludes particular routine tools.

A vendor calling a product “analytics” instead of “artificial intelligence” does not automatically answer whether a system falls within the law. Conversely, not all uses of an ordinary database necessarily trigger the same obligations.

Examples worth reviewing include applicant-ranking systems, attendance or productivity scoring, automated performance recommendations, worker deactivation systems and algorithm-assisted discipline. A careful audit must consider how the business actually uses the tool, not merely the product label.

SB 951: additional information when AI contributes to covered layoffs

Another measure signed September 30, 2026, SB 951 (Chapter 860), modifies California’s Worker Adjustment and Retraining Notification framework for certain technology-related displacement.

Where an employer is already required to provide a covered Cal-WARN notice and a qualifying mass layoff, relocation or termination is caused in whole or in substantial part by AI or other automation replacing or automating jobs, the law calls for additional details about technology displacement. The obligation is not the same as a new 60-day notice requirement for every small business that adopts software.

Employers contemplating substantial operational restructuring should assess Cal-WARN coverage, timing, which roles are affected, whether automation is a substantial cause and which information must appear in the notice. The applicable statutory language and effective dates should be reviewed with counsel before decisions are announced.

AI-enabled workplace surveillance raises separate issues

California also enacted AB 1883, addressing certain AI-enabled workplace surveillance practices beginning January 1, 2027, subject to its scope and exceptions. It restricts using covered tools to infer or recognize an employee’s emotional state or collect specified neural data in workplace contexts.

A company should not assume a monitoring product is legally harmless because it was marketed as a wellness or productivity feature. The data collected, the employer’s purpose, employee notice and applicable privacy and employment rules should be considered before deployment.

This topic also touches other legal duties, including California privacy law and sector-specific requirements; the existence of an AI-employment statute does not displace the need to assess those obligations.

A practical 2027 checklist for California employers

Rather than starting with a blanket “stop using AI” directive, businesses can begin with an inventory and risk-based review:

  • Inventory all systems used for recruitment, screening, scheduling, performance reviews, discipline, termination and worker monitoring.
  • Map decision points: identify where a system merely displays information versus materially influencing decisions that affect workers.
  • Identify the accountable human: for covered decisions, document genuine review of relevant evidence and the authority to depart from the automated recommendation.
  • Assess discrimination risks: examine how data, scoring criteria and vendor models could disadvantage protected groups.
  • Review record retention: confirm whether required records and automated-decision data can be retained and produced when legally necessary.
  • Review employee communications: prepare notices and data-response procedures where the new statutes require them.
  • Check vendor agreements: address information access, audit cooperation, confidentiality, security, technical changes and indemnification without assuming a vendor contract alone establishes compliance.
  • Coordinate restructuring: if automation may lead to substantial workforce changes, obtain advice on Cal-WARN and other employment obligations before announcing a plan.

These steps are operational controls, not a claim that a single checklist guarantees compliance. Employers should confirm which laws apply to their size, type, workforce and actual technology deployment.

A hypothetical decision that illustrates the risk

A Pasadena employer uses a service that assigns productivity scores and automatically recommends termination for workers below a numerical threshold. Supervisors routinely accept the recommendations without looking at the underlying records. The company considers that “human approval” because a supervisor clicks the final button.

Under a framework requiring genuine human corroboration for certain covered decisions, that practice may not be enough. The relevant questions would include how the system generated its score, what evidence the supervisor independently reviewed, whether legally required notices were provided and whether the criteria created unlawful discrimination or retaliation risks.

The example illustrates a compliance concern, not a finding about a real business.

Prepare your systems and contracts before the deadline

The best time to review an automated decision-making workflow is before a contested termination, a discrimination complaint or a rushed vendor rollout. Business owners need a documented understanding of what their tools do, who makes the consequential decisions and which legal duties apply.

For related business matters, read California AB5 Independent Contractor Rules and California Non-Compete Agreement Ban. To discuss contractual or general counsel questions affecting your business, contact Lum Law Group.

Primary sources: California Civil Rights Department automated-decision regulations; SB 947 bill record; SB 951 bill record; Governor’s 2026 signing announcement.

Editorial: Draft prepared November 17, 2026; attorney verification required before publication, including all enacted effective dates, covered entities and exceptions. General information only, not individualized employment or business-law advice.

Is Your California Business Ready for 2027?

Learn how the relevant legal requirements may affect your situation. A consultation is an opportunity to discuss your specific questions and next steps.

Frequently Asked Questions

Is all AI use in California workplaces prohibited in 2027?

No. The statutes and regulations impose specified restrictions and duties on covered systems and decisions; they do not categorically ban AI in the workplace.

The relevant new automated-decision provisions are scheduled to become operative July 1, 2027. Employers should verify the enacted text and any subsequent amendments or agency guidance.

Yes. Civil Rights Department regulations clarifying antidiscrimination requirements for automated employment systems took effect October 1, 2025.

No. An employer should examine its own use of the system, the applicable obligations and what the vendor contract and technology actually provide.

Not as a universal rule. Its relevant provisions operate in the context of covered California WARN events and specific statutory conditions; a separate legal review may be needed for the particular workforce change.

SHARE THIS ARTICLE

About A. Justin Lum

A. Justin Lum has been practicing law since 1993, with more than 30 years of experience serving clients across a wide range of legal matters. A graduate of the University of California, Berkeley and UC Davis School of Law, he is also a past President of the Southern California Chinese Lawyers Association, continuing a proud family legacy of leadership in the legal profession.

Legal Perspective

For an employer using automated systems, legal review should focus on how the technology affects real employment decisions, whether human oversight is meaningful, what records are retained and which notices the specific law requires. General information for attorney review, not an attorney quotation.