California’s Expanded Fair Employment Protections Under Automated Decision Systems
The California Civil Rights Council has clarified that the state’s Fair Employment and Housing Act (FEHA) applies to automated decision systems (ADS) used in hiring and employment management. This ruling explicitly extends anti-discrimination protections to AI-driven tools, ensuring compliance with California’s civil rights laws regardless of whether the decision-making process is human or algorithmic. The regulation takes effect on October 1, 2025, marking a significant shift in how employers and vendors must approach the use of AI in workforce-related decisions.
The Civil Rights Council, a seven-member volunteer body appointed by the Governor and confirmed by the Senate, oversees the interpretation and enforcement of these regulations. Established in 2013 as a successor to the Fair Employment and Housing Commission, the Council operates under the California Civil Rights Department (CRD). Its role includes promulgating regulations that interpret California’s civil rights laws, including those now addressing ADS in employment contexts.
The Council’s authority extends to creating technical assistance for civil rights advisory bodies across the state, ensuring that employers, vendors, and other stakeholders understand their obligations under FEHA. All Council activities are conducted openly, with meetings, agendas, and reports accessible to the public. Interested parties can subscribe to the Council’s mailing list for updates on meetings, notices, and regulatory developments, though subscribers must configure their email systems to accept messages from the state’s LISTSERV domain.
Key Regulatory Requirements for Automated Decision Systems
One of the most critical provisions of the new regulation is the potential liability of vendors supplying ADS tools. Under the updated rules, vendors may be held legally accountable as agents if their systems contribute to discriminatory outcomes in hiring or employment management.
Employers and vendors must retain data related to ADS operations for a minimum of four years.
Employers must ensure that their AI tools do not perpetuate bias or discrimination, whether through flawed algorithms, biased training data, or inadequate oversight.
Implications for Employers, Vendors, and the AI Industry
For employers in California, the regulation introduces new compliance obligations that extend beyond traditional hiring practices. Companies using ADS for recruitment, promotions, or workforce management must now assess whether their tools align with FEHA’s anti-discrimination principles.
Vendors developing or selling ADS tools must also adapt to the heightened scrutiny under California law. The potential for vendor liability means that companies offering AI-driven hiring solutions will need to implement robust safeguards, such as bias mitigation strategies and transparent documentation of system behavior.
The regulation’s impact extends beyond California, as it sets a precedent for how other states and jurisdictions may approach AI governance in employment contexts. While the law applies only to entities operating within California, its influence could encourage similar measures nationwide, particularly in regions with strong labor protections or growing concerns about algorithmic discrimination.
Next Steps for Compliance and Enforcement
Employers and vendors should begin reviewing their current ADS practices to identify gaps in compliance with FEHA. This may involve conducting internal audits of AI tools, updating vendor contracts to include liability clauses, and establishing data retention policies that meet the four-year requirement. The Civil Rights Department’s resources, including technical assistance and public notices, can provide further guidance on implementing these changes.
As the regulation takes full effect, stakeholders should monitor updates from the Civil Rights Council, including any additional guidance or enforcement actions. The Council’s public meetings and reports will offer insights into how the law is being interpreted and applied in practice. Proactive engagement with these resources can help organizations avoid compliance risks and contribute to a more equitable use of AI in employment.
Required actions
- Do not use automated-decision systems that discriminate against applicants or employees on protected grounds; anti-bias testing is relevant to liability - Vendors acting as the employer's agent can be liable - Keep ADS-related records, including data and outputs, for four years