California AB 2930: AI Hiring Law — Complete Employer Guide (2026)

Citation: California AB 2930 (Government Code § 12153 et seq.) Effective: January 1, 2026 Jurisdiction: California — applies to all employers operating in California or hiring for California positions Enforced by: California Attorney General; California Civil Rights Department Official Source: California Legislative Information – AB 2930

California's AI hiring law is in effect. If your company uses any software to screen resumes, score video interviews, rank candidates, or generate hiring recommendations in 2026 — and those candidates include anyone in California — you are subject to AB 2930. The law imposes strict obligations around pre-use disclosure, annual bias testing, data minimization, and an explicit right to opt out and request human review. Penalties reach $7,500 per violation, and each affected candidate counts as a separate violation.

This guide covers every obligation California employers face, explains exactly what qualifies as an "automated decision system," and gives you a practical compliance roadmap you can act on today. For a side-by-side comparison of every state AI hiring law, see AI hiring laws by state. If you need to assess your current exposure, start with our free compliance assessment.

What Is California AB 2930?

California AB 2930, signed into law in 2024 and effective January 1, 2026, is California's first comprehensive statute regulating the use of automated decision systems (ADS) in employment. It establishes four interconnected obligations: transparency (tell candidates an ADS is being used), fairness testing (prove it doesn't discriminate), data restraint (collect only what you need), and human oversight (honor requests for a human reviewer).

The law builds on existing California employment protections — including the Fair Employment and Housing Act (Government Code § 12940 et seq.) and the California Consumer Privacy Act / California Privacy Rights Act (Civil Code § 1798.100 et seq.) — but adds a new layer of affirmative obligations specific to AI-driven hiring processes. Unlike those frameworks, AB 2930 doesn't require a complaint to trigger obligations: employers must act proactively, before a candidate ever raises a concern.

The law was modeled in part on New York City Local Law 144 and Colorado SB24-205, but it goes further on data minimization and individual opt-out rights than either predecessor. For Illinois employers with multi-state hiring operations, note that Illinois AIVIA (820 ILCS 42) has parallel but distinct video interview requirements that may apply to the same candidates.

What Is an "Automated Decision System" Under AB 2930?

AB 2930 defines an automated decision system (ADS) as any computational process — including machine learning, statistics, or other data-processing techniques — that generates a decision, recommendation, score, classification, or output that influences a decision affecting a job applicant or current employee.

The definition is deliberately broad. If a vendor's software produces a score, a ranking, a flag, or even a "review suggested" tag that a recruiter relies on — that's an ADS. The law does not require that the AI make the final decision; influencing it is enough.

Tools Covered Under AB 2930

Vendor responsibility: Employers remain responsible for ensuring their vendors' tools comply. Obtain compliance representations in writing and build AB 2930 obligations into vendor contracts. The EEOC's April 2023 guidance on AI hiring and the FTC's guidelines on AI both signal that vendor-supplied tools don't insulate employers from liability.

The Four Core Obligations

1. Pre-Use Disclosure (Notice Requirement)

Before an ADS is used to evaluate any California applicant, the employer must provide clear, prominent, plain-language notice that includes:

• The fact that an ADS will be used in the hiring process
• The general purpose of the ADS (e.g., "to evaluate whether your skills match the role's requirements")
• The categories of data the ADS collects and processes
• Contact information for submitting a human review request or asking questions

What doesn't count: A disclosure buried in paragraph 14 of a terms-of-service document does not satisfy AB 2930. The notice must be prominent — ideally on the job application page itself, in the initial candidate communication, or in a clearly labeled "Hiring Process" section that candidates see before they begin any ADS-evaluated step.

Practically, this means updating your California job application portals, careers page, and any pre-interview communications. If you use third-party ATS platforms like Workday, Greenhouse, or Lever, you need to implement disclosure messaging at the point where ADS tools are first engaged. Our AI Hiring Disclosures tool generates compliant notice language for every major ATS format.

2. Annual Bias Testing

Every ADS used in California hiring must undergo an annual bias evaluation designed to identify whether the tool produces disparate impact across legally protected categories, including:

• Race, color, and national origin
• Sex and gender identity
• Religion
• Age (applicable to candidates 40+, consistent with the ADEA, 29 U.S.C. § 623)
• Disability status
• Any other class protected under the Fair Employment and Housing Act

What the testing must cover:

• Documented testing methodology (consistent with or more rigorous than the Uniform Guidelines on Employee Selection Procedures, 29 C.F.R. § 1607)
• Results retained for at least three years
• If disparate impact is found: documented corrective action — modification, reconfiguration, or discontinuation of the ADS — within a reasonable timeframe

Note that AB 2930 permits internal testing by the employer with a documented methodology. This differs from NYC Local Law 144, which requires an independent third-party bias audit and public publication of results. Colorado's law (Colo. Rev. Stat. § 6-1-1701) similarly requires impact assessments but also mandates annual reporting to the Colorado AG — an obligation AB 2930 does not impose.

3. Data Minimization

AB 2930 introduces a data minimization principle that directly limits what ADS tools can collect and process about California candidates. The ADS may only use candidate data that is directly relevant to the specific position being filled.

This means:

• Social media monitoring unrelated to job requirements is prohibited
• Health, biometric, or family status data that doesn't bear on job performance cannot be fed into the ADS
• Employers must review vendor data collection practices and confirm only job-relevant signals are used

Practical action: Request a data inventory from every ADS vendor you use for California hiring. If a vendor cannot tell you exactly what inputs their model uses, that is a compliance risk. Data minimization under AB 2930 intersects with the California Consumer Privacy Act (Civil Code § 1798.100) — candidates have CCPA rights over their ADS-processed data, including the right to know, the right to delete, and the right to opt out of certain data uses.

4. The Right to Opt Out and Request Human Review

California AB 2930 grants every applicant the explicit right to opt out of ADS evaluation and request that a human reviewer independently assess their candidacy. This is the most operationally significant obligation for employers.

Requirements:

• The opt-out right must be disclosed as part of the pre-use notice
• Employers must establish and communicate a clear process for submitting requests
• Human reviewers must conduct an independent assessment — they cannot simply review the AI's output and endorse it
• Candidates cannot be penalized, rejected, or treated differently for exercising this right

Building a compliant human review workflow requires cross-team coordination: HR must be trained to recognize opt-out requests, recruiters must know how to conduct AI-free reviews, and your ATS must flag candidates who have exercised the right so they're tracked separately. See our AI Hiring Compliance Training module for onboarding templates.

AB 2930 vs. Other State AI Hiring Laws

California's law is one of the most comprehensive in the country, but it sits alongside a growing patchwork of state and local regulations. Here's how it compares to the laws most likely to affect multi-state employers:

For a full comparison across all states, see the AI hiring laws by state tracker. California employers with New York City offices also need to track NYC LL144 compliance separately — the two regimes overlap for candidates evaluated for NYC roles.

Penalties and Enforcement

Each candidate whose rights are violated is a separate violation. For employers running high-volume AI-screened hiring in California — thousands of applicants per quarter — aggregate liability from a single un-disclosed ADS can reach seven figures.

The California AG and Civil Rights Department have signaled that AI-related employment enforcement is a priority in 2026. Federal frameworks add further exposure: the EEOC's guidance on AI hiring tools makes clear that Title VII (42 U.S.C. § 2000e) liability attaches to AI tools that produce disparate impact — even where the employer did not intend to discriminate. The Department of Labor's OFCCP has also indicated it will scrutinize AI use in federal contractor hiring under existing adverse impact standards.

2026 Compliance Roadmap

Immediate actions (if not already done):

• Inventory every ADS tool used in California hiring — including vendor-supplied features inside your ATS
• Request written compliance representations from all ADS vendors
• Draft pre-use disclosure notices in plain language and add to California job application portals
• Build a human review request process and train HR to handle it
• Update California CCPA/CPRA privacy notices to disclose ADS data collection and candidate rights

Q3 2026 and ongoing:

• Complete first annual bias testing cycle; document methodology and results
• Establish 3-year record retention for all bias testing documentation
• Conduct data minimization review — confirm each ADS uses only job-relevant inputs
• Audit vendor contracts to ensure AB 2930 compliance obligations flow down
• Train recruiters on what opt-out requests trigger and how to conduct AI-free reviews

Use our AI Hiring Compliance Checklist to walk through every obligation step-by-step with documentation prompts.

How This Connects to Federal Law

AB 2930 operates alongside — not instead of — existing federal frameworks. California employers must satisfy all of the following simultaneously:

• Title VII of the Civil Rights Act (42 U.S.C. § 2000e): Prohibits AI tools that produce disparate impact against protected classes. The EEOC's Uniform Guidelines on Employee Selection Procedures (29 C.F.R. § 1607) govern how employers validate selection tools.
• ADEA (29 U.S.C. § 623): Protects workers 40+ from age discrimination — AI tools that de-prioritize older applicants face ADEA exposure even where not intentional.
• ADA (42 U.S.C. § 12112): AI tools that screen out candidates based on disability-correlated proxies (e.g., employment gaps, non-linear career histories) may constitute discriminatory screening.
• FCRA (15 U.S.C. § 1681): If your ADS incorporates background check data or credit information, FCRA obligations apply. The CFPB has issued guidance on AI-assisted background screening.

The EEOC's enforcement guidance and DOJ IER resources on automated hiring are the primary federal reference points alongside AB 2930.

Frequently Asked Questions

Does AB 2930 apply to fully remote roles where the candidate is in California?

Yes — and this is one of the most frequently misunderstood aspects of the law. If a candidate is located in California and you use an ADS to evaluate them, AB 2930 applies regardless of where your company is headquartered, where the job is physically located, or whether the role is remote. An employer based in Texas that hires a remote engineer in San Francisco is subject to AB 2930 for that hire. This extraterritorial reach is consistent with how California generally applies its employment laws and mirrors the approach taken under the CCPA/CPRA.

Practical implication: If you have any California-based employees or remote candidates, your entire ADS compliance stack needs to meet California standards — you can't maintain a separate California-only configuration and a non-compliant default.

What's the difference between AB 2930 bias testing and the independent audit required under NYC Local Law 144?

They have the same goal — detecting discriminatory bias — but NYC's standard is significantly more rigorous. NYC Local Law 144 (NYC Admin. Code § 20-871) requires an independent third-party auditor to conduct the bias evaluation, and employers must publish the results on their website in a publicly accessible format. AB 2930 permits internal testing by the employer as long as the methodology is documented and results are retained for three years. Internal testing is less expensive and logistically simpler, but the documentation burden is real: you must be able to demonstrate to the California AG exactly what you tested, how you tested it, and what you found. For multi-state employers, we recommend designing your bias testing program to meet NYC's independent audit standard — it satisfies both regimes and produces stronger defensible documentation.

Can we continue using an ADS tool while we're in the process of getting it bias-tested?

This is a practical gray area the law doesn't address explicitly. The statute requires annual bias testing, but it doesn't specify that testing must be complete before initial deployment. However, using an ADS that you haven't tested creates significant liability exposure under both AB 2930 and Title VII (42 U.S.C. § 2000e) if the tool later proves to have disparate impact. Best practice: complete bias testing before deploying any new ADS in California hiring, and prioritize testing for tools already in use. Document your testing timeline as part of your compliance record. If a complaint arises while testing is in progress, your documented diligence is your best defense.

What qualifies as a valid "human review" when a candidate opts out of AI evaluation?

A valid human review under AB 2930 is an independent assessment — not a review of the AI's output. A recruiter who sees the AI's score and then endorses it has not conducted an independent review. Instead, the human reviewer must evaluate the candidate's application, resume, and qualifications without reference to ADS outputs, using the same criteria they would apply in a non-AI hiring process. This means your HR team needs a documented, AI-free review workflow that produces a separate record for each opt-out candidate. In practice, this typically means: (1) flagging opt-out candidates in your ATS, (2) routing them to a designated human reviewer pool, (3) evaluating them against the same job criteria without ADS scores, and (4) documenting the review outcome separately. Candidates must be informed of the outcome of their human review.

What does "data minimization" actually prohibit in practice?

AB 2930's data minimization principle prohibits ADS tools from using candidate data that isn't directly relevant to the specific position. In practice, this means: the ADS cannot incorporate social media content unrelated to job performance, infer personality traits from off-topic behavioral signals, or process sensitive personal information like health status, family composition, or financial data that doesn't predict job-relevant capability. It also means an ADS tool calibrated for one job type cannot be applied to a fundamentally different role without retesting — the "relevance" requirement is position-specific, not employer-wide. Request a full data dictionary from every ADS vendor you work with. If a vendor can't or won't tell you exactly what inputs their model uses, that's a red flag that triggers both AB 2930 data minimization concerns and your due diligence obligations under California law.

Does AB 2930 cover internal promotions and performance evaluations, not just external hiring?

Yes. AB 2930 covers "employment decisions" broadly, which includes promotion, transfer, demotion, and other employment actions — not just initial hiring. If your company uses AI tools to identify high-potential employees for promotion tracks, or to generate performance ratings that feed into compensation decisions, those tools are within scope for California employees. This is a broader reach than NYC Local Law 144, which is largely focused on hiring. Make sure your ADS inventory includes internal HR analytics platforms and performance management tools used for California-based employees. See our AI governance statement generator for help documenting your policy across the full employment lifecycle.

Can we satisfy the opt-out right by simply removing the candidate from AI-screening and advancing them to a human stage?

Partially — but the process must be genuinely equivalent. Advancing an opt-out candidate to a human review stage is the right structural approach, but the candidate cannot be disadvantaged in the process. They must be evaluated against the same job criteria, given the same opportunities to demonstrate qualification, and not face any delay or deprioritization relative to AI-screened candidates simply because they exercised their opt-out right. If your hiring pipeline is designed such that AI-screened candidates move to interviews within three days while opt-out candidates wait in a manual queue for two weeks, that disparity likely violates the non-retaliation principle built into AB 2930. Build your opt-out workflow with the same SLAs as your standard pipeline.

How does AB 2930 interact with the California Consumer Privacy Act (CCPA)?

The two laws overlap significantly for ADS-processed candidate data. Under the CCPA (Civil Code § 1798.100), California job applicants have the right to know what personal data you've collected about them, the right to request deletion of their data, and the right to opt out of the sale or sharing of their personal data. AB 2930 extends these rights specifically to ADS-generated data: candidates have the right to know what the ADS collected, what categories of data influenced its output, and contact information for follow-up. Your California privacy notice must reflect both sets of rights. Practically, this means one coordinated update to your privacy notice and candidate data FAQ rather than two parallel disclosures — but both must be addressed. For CCPA-specific compliance, see the CFPB's guidance on automated screening and consumer rights.

California Official Resources

• California Legislative Information – AB 2930 full text
• EEOC: Questions and Answers on AI Hiring and Selection Procedures
• DOJ Immigrant and Employee Rights Section – AI and Employment Discrimination
• CFPB: Consumer Protection and Automated Background Screening
• OFCCP: Federal Contractor AI Hiring Compliance

Related Laws and EmployArmor Resources

• Colorado AI Act (SB24-205) — Full Guide
• NYC Local Law 144 — Independent Audit Requirements
• Illinois AIVIA — Video Interview AI Rules
• New Jersey S-1943 — Automated Employment Decision Tools
• AI Hiring Laws by State — 2026 Tracker
• AI Hiring Compliance Checklist
• AI Hiring Disclosures Tool
• Bias Audit Coordination Tool
• AI Hiring Compliance Training
• Free Compliance Assessment — See Your California Exposure

Last updated: June 2026. This content is for informational purposes only and does not constitute legal advice. Consult qualified California employment counsel for guidance specific to your organization and hiring practices.