California FEHA AI Rules - Solution

Overview

California’s FEHA rules on AI in employment

New updates to California’s Fair Employment and Housing Act (FEHA) go into effect on October 1, 2025. These changes introduce direct legal responsibilities for both employers and HR Tech vendors using Automated Decision Systems (ADS) in employment.

Automated Decision Systems or ADS in employment is technology that helps make or influence employment-related decisions, using algorithms or artificial intelligence (AI). These systems are increasingly used in areas such as:

Screening job applications (e.g., resume parsers or automated filtering tools)

Interview analysis (e.g., AI-powered video interview scoring)

Hiring recommendations

Promotion or performance evaluations

Termination or workforce reduction tools

Compliance

The rules

These rules are among the most comprehensive state-level AI employment regulations in the US and sets a precedent for how AI in employment is used.

Who the law applies to

An employer using AI or ADS technology in hiring or promotion

A vendor developing AI for employment

A compliance or legal lead managing AI risk

Key requirements - FEHA’s ADS amendments outline specific requirements:

Notice
‍Candidates and employees must be notified when an ADS tool is used to make employment decisions

Bias testing
‍Is not mandated, but courts and regulators will treat the quality, scope, and recency of testing as central in deciding liability

Record keeping
All testing results and design documentation must be retained for at least 4 years

Safeguards
Employers and vendors must implement reasonable safeguards to prevent unlawful discrimination

Joint liability
Vendors who provide ADS tools may be held jointly liable with employers for discriminatory outcomes

Reducing risk

Both the buyers and builders of AI tech used in employment decisions are responsible for actively reducing the risk of discrimination. If you're using AI to help with hiring, promotions, or firing - you can’t just assume it’s fair. You have to take reasonable, proactive steps to make sure your system isn't treating people unfairly based on protected characteristics, this includes:

Regular bias testing across all protected groups (e.g., by race, gender, age)
Keeping clear records of how your AI works and how it's been tested
Giving people notice when AI is used to assess them
Giving humans the final say - especially in high-stakes decisions
Using independent audits (like Warden AI) to verify fairness

Protected Classes

Protected characteristics

FEHA prohibits discrimination across 18 protected classes in California employment law, including:

Race
Religious creed
Colour
National origin
Ancestry
Physical disability

Mental disability
Medical condition
Genetic information
Marital status
Sex
Gender

Gender identity
Gender expression
Age
Sexual orientation
Reproductive health decision making
Military and veteran status

Preparation

How Warden AI gets companies ready for The California FEHA Regulation

Warden AI provides independent auditing and certification of AI employment systems, helping both vendors and employers prepare for compliance:

Compliance support

Warden’s platform provides bias audits, documentation trails, and automated reporting aligned with regulatory requirements.

Testing across all applicable protected classes

Our audits are built on robust methodology covering up to 12 FEHA-protected characteristics.

Audit trails

Maintain a time-stamped record of system behaviour - whether you're in pre-market testing or in production.

Dashboards and reporting

Share testing results and risk information with stakeholders - and demonstrate safeguards in place.

Join Our Webinar

California’s New Rules on AI in Employment: What Employers and Vendors Need to Know

Wednesday, September 24th, 1 PM EDT | 10 AM PDT | 6 PM BST

Together with Maneesha Mithal (Partner at Wilson Sonsini Goodrich & Rosatim, previously The FTC) we will be covering California’s New FEHA rules on AI in employment.

Be compliant for California’s FEHA rules