Recent legislative and regulatory developments in California are reshaping how employers may use Automated Decision Systems (ADS), including Artificial Intelligence (AI), in employment-related decisions. The California Civil Rights Council (CRC) has finalized regulations under the Fair Employment and Housing Act (FEHA) to address potential discrimination risks and clarify how the law applies in this evolving technological landscape. These regulations took effect on October 1, 2025. In addition, Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act, strengthens employee protections by establishing whistleblower safeguards for individuals who report safety risks associated with large AI models. The law also requires large AI developers to implement secure, anonymous internal reporting channels and prohibits retaliation against employees who make such disclosures.
Clarifying Automated-Decision Systems (ADS)
The regulations apply to all California employers with five or more employees and define an ADS broadly.
- Definition. An ADS is a “computational process that makes a decision or facilitates human decision making regarding an employment benefit”.
- Scope. This includes systems that use AI, machine-learning, algorithms, statistics, or other data processing techniques.
Examples of Use
ADS can include tools that:
- Screen resumes for particular terms or patterns.
- Direct job advertisements to targeted groups.
- Assess applicants’ skills through computer-based tests, games, or challenges.
- Analyze facial expressions or voice in online interviews.
Key Compliance Requirements for Employers
The core of the new regulations is the prohibition against using an ADS or selection criteria that results in discrimination against applicants or employees based on characteristics protected under FEHA.
- Unlawful Discrimination. An employer may be liable for unlawful discrimination, even if unintended, if the use of an ADS results in a disparate impact on a protected class.
- Unlawful Medical Inquiries. ADS assessments, such as tests or puzzles, which are likely to elicit information about a disability may be considered an unlawful medical or psychological inquiry.
- Third-Party Liability. Employers may be held liable even if the ADS is developed or administered by a third-party vendor. The regulations expand the definition of an “agent” to include third parties that perform traditional employer functions using an ADS.
- Recordkeeping. Employers must now preserve personnel and other employment records, including ADS-related data, for a minimum of four years, an increase from the previous two-year requirement.
- Anti-Bias Testing. Although not strictly mandated, evidence of anti-bias testing or similar proactive efforts to avoid unlawful discrimination can be used in defense of a discrimination claim. Conversely, the absence of such efforts may be used against an employer.
Next Steps for Employers
The regulations took effect on October 1, 2025, employers who have yet to take proactive steps should ensure their use of automated systems is compliant and minimizes risk:
- Inventory Your Tools: Audit all automated technologies currently used in human resource related functions, including recruitment, hiring, promotions, training, and evaluation, to confirm if they fall under the broad ADS definition.
- Update Policies & Procedures: Review and update your record retention policies to ensure all ADS-related data and employment records are preserved for the mandated four-year period.
- Vendor Management: Establish guidelines for managing relationships with third-party ADS vendors. Ask vendors about their anti-bias testing protocols, data use practices, and clarify their understanding of their potential liability as your agent.
- Consider Bias Testing: Establish a plan for conducting bias audits of your ADS tools, documenting the criteria, results, and any steps taken to address potential discriminatory outcomes.
