As the rise of artificial intelligence (AI) continues to pose threats while still unchecked, California legislators and civil rights groups are moving to regulate AI use in making employment decisions to counteract possible discrimination.

Currently, The Golden State is cracking down on automated employment bias through its two-pronged approach, which involves the California Civil Rights Council’s (CCRC) proposed amendments to the Fair Employment and Housing Act (FEHA) and Assembly Bill 2930.

Nevertheless, these significant steps are not only taken in California since other U.S. states are rushing to create AI-related laws and regulations to address the societal dangers that the rapidly advancing technology may bring.

Proposed FEHA Amendments

FEHA was passed into law on September 18, 1959. At the time of its legalization, it served as a protection against sexual harassment and other unlawful employment and housing discrimination. Today, it needs to do more—to protect the workforce against the potential harm of AI.

Sharing the same sentiment, the CCRC held a public hearing at the University of California Berkeley School of Law on July 18 to discuss the proposed amendments to the act.

In the hearing, the public raised the need to improve the definitions and scope of the proposed rules, which included employment agencies and automated decision systems.

Considering the written and oral testimonies, the proposed changes indicate that the revised FEHA will affect those who regularly pay at least five individuals for work or services, as well as employers’ agents and employment agencies. The CCRC stated that this broad definition aims to spread the coverage and accountability for biased practices resulting from the use of AI in employment.

Further, the amendments define an “automated decision system” as a computational process that screens, evaluates, categorizes, recommends, makes, and facilitates decisions that impact employees or applicants.

Systems that utilize machine learning, algorithms, statistics, or other AI techniques in performing computer-based tests, targeted job advertisements, resume screening, online interviews, and similar activities also fall within the definition.

These proposed rules, which also cover employer impact, consideration of criminal history, and record-keeping responsibilities, are intended to update employment practices in California and follow existing procedures, particularly the White House’s Blueprint for an AI Bill of Rights and the Equal Employment Opportunity Commission’s guidelines on algorithmic fairness.

AB 2930

Another bill that seeks to fight against algorithmic discrimination caused by the use of AI is AB 2930.

Under this bill, “algorithmic discrimination” is defined as the unjust disadvantage put to particular people based on their gender, sexual orientation, race, and other characteristics in hiring, termination, and salary.

“We see example after example in the AI space where outputs are biased,” noted California State Representative Rebecca Bauer-Kahan, one of the bill’s supporters.

AB 2930 was first drafted in 2023 but failed due to opposition from major tech companies. In 2024, the anti-discrimination bill was reintroduced, and it is now under the Appropriations Committee after passing the Senate’s Judiciary Committee in July.

This year’s reintroduction saw a turn as AI titans Microsoft and Workday expressed support for the bill. However, this proved to be short-lived after they raised concerns over proposed changes that would require them to develop their products to minimize bias.

Meanwhile, debates surrounding AB 2930 remain, with one side insisting that the bill requires stronger language and the other saying it is too broad.

Discussions regarding the bill are expected no earlier than August 5, when the Legislature returns from Summer Recess. The decision to finalize AB 2930 should be made before the end of August.