On October 1, 2025, an updated regulations from the California Council of Civil Rights formally limit the use of artificial intelligence (AI) tools by California employers in employment decision-making. In employment situations, these tools can be applied in a variety of ways to manage the workforce, including screening resumes, forecasting about applicants or employees, measuring applicants or employees' skills or capabilities, hiring directly employment advertisements, recruitment and/or recommendations to target groups, and/or recommendations, and/or recommendations.
New rules apply existing anti-discrimination laws to AI tools
Rather than devising an entirely new set of regulations on AI, the Civil Rights Council has decided to amend the existing regulatory framework of the Fair Employment and Housing Act (FEHA) into an “automatic decision system.” Therefore, AI regulations function as another source of potentially unacceptable discrimination based on protected characteristics. Normally such tactics would not be ignorant as Caselaw interprets Feha is well established, but in this example the latest regulations of the Civil Rights Council presents important changes such as:
- For the first time, an employer's “agent” as someone acting directly or indirectly on behalf of the employer, “exercises functions traditionally exercised by an employer or other FEHA regulatory activity.” The regulations specifically identifies decisions regarding the applicant's recruitment, review, employment, promotion, or salary, benefits, or leave as potential examples of such “functions” and specify that an automated decision system is eligible as an employer's agent.
- An automated decision-making system is broadly defined as “.[a] A calculation process that determines or facilitates human decisions regarding employment benefits. The updated regulations consider automated decision-making systems, along with AI, along with “machine learning, algorithms, statistics, and/or other data processing methods.”
- We define AI as “a machine-based system that infers from the inputs that receive and how it generates outputs.” do not have It's a generally accepted explanation of AI, but it's much closer to the definition of a computer instead. As drafted now, this very broad definition of AI could include computer programs within range, even if such programs are not performing the complex tasks normally performed by the normal season.
- Establish new, clear positive defenses for employers using automated decision-making systems in FEHA lawsuits: evidence of anti-bias testing or similar efforts to avoid illegal discrimination.
- Employers require that you keep a record of decision data from your automated decision system for at least four years.
The updated rules do not cover explicit bias only. These also apply to algorithms that negatively affect protected groups, often referred to as “differential influences.” Therefore, even if employers do not intentionally use AI to distinguish between applicants or employees, the new rules prohibit the use of AI that has different effects.
Employers have multiple options to mitigate the risk of violating the updated Civil Rights Board regulations
October 1, 2025, the updated regulations effective date will provide several months to California employers to ensure that the use of AI or any computer system is in compliance with the law in the context of workforce management. In light of the formal perception of anti-bias training positive defenses with updated regulations, employers should consider trying to reduce the risk that AI use has different effects based on protected characteristics. As recognized in the updated rules, there are various ways that risk reduction can have an impact, such as:
- Perform bias audits or impact assessments before implementing AI technology.
- Regularly check the impact of AI tools on post-implementation protection groups.
- Evaluation of the quality of audits, assessments, and/or tests used to detect potential discriminatory outcomes in the use of AI tools.
- Document the results of the anti-bias test, including employer responses to the results. and
- Enact human reviews of employment decisions promoted by AI.
New rules: Impact on labor management vendors
As updated regulations broadly define employer agents, vendor management plays a key role in mitigating the risks associated with the use of AI, as the owner or licensor of an automated decision-making system may be certified as an employer agent under FEHA. California employers should carefully consider compensation and defense contracts with vendors, but they also require vendors to prove the effectiveness and results of anti-bias tests implemented on AI platforms. The use of AI in workforce management presents exciting opportunities for California employers, but such use is risky if you don't look diligently towards legal compliance and the right risk balance.
