In June 2026, a federal judge overseeing the closely watched Mobley v. Workday case signaled that she would allow California discrimination claims against the software company to move forward by largely denying Workday’s motion to dismiss. U.S. District Judge Rita Lin rejected arguments that state anti-discrimination laws should not reach hiring decisions made through an out-of-state platform.
The case, which centers on allegations that Workday’s AI-powered screening tools disadvantaged applicants based on race, age, and disability, has become a proving ground for a question California regulators already answered on their own: when an algorithm makes or influences an employment decision, the employer, and often the vendor behind the tool, can be held responsible if that decision discriminates.
That answer is now incorporated in the Fair Employment and Housing Act (FEHA) itself. Since October 1, 2025, new regulations from California’s Civil Rights Council have established that using artificial intelligence or an automated decision system does not change who is liable when a hiring, promotion, or termination decision turns out to be discriminatory.
What Counts as an Automated Decision System Under FEHA
The regulations define an Automated Decision System, or ADS, broadly: any computational process, including one built on artificial intelligence, machine learning, algorithms, or other data processing techniques, that makes a decision or helps a human make one regarding an employment benefit. That definition sweeps in far more than resume-screening software. Under the regulations, an ADS can include:
- Tools that screen resumes or applications for particular terms, keywords, or patterns;
- Video interview platforms that analyze a candidate’s facial expressions, word choice, or vocal tone;
- Computer-based tests, games, or puzzles used to predict skill, aptitude, or reaction time;
- Systems that direct job advertisements to specific demographic groups while excluding others; or
- Productivity or behavioral scoring tools used to evaluate current employees for discipline, promotion, or termination.
Critically, the rules apply even when a human makes the final call. If an algorithm narrows the applicant pool, ranks candidates, or flags an employee for review, and that output shapes or influences a subsequent human decision, the ADS is part of the legal analysis. An employer does not avoid liability simply because a person clicked the final button.
Employers Cannot Point to Their Vendor to Avoid Liability
One of the most significant changes in the regulations is how they treat third-party vendors. Many employers using AI hiring tools did not build those tools themselves, they licensed them from outside companies that specialize in applicant screening, video assessment, or workforce analytics. The regulations close off the argument that responsibility for a discriminatory outcome belongs solely to the vendor.
Under the amended rules, a vendor performing recruitment, screening, or similar functions on an employer’s behalf can be treated as the employer’s agent for purposes of FEHA liability. That means both the employer that deployed the tool and the company that built it may face exposure when the tool produces discriminatory results.
This mirrors the theory that has kept the Workday litigation alive: courts have found it plausible that a company providing algorithmic hiring services to thousands of employers can be treated as each of those employers’ agent, exposing the platform itself to direct liability alongside its customers.
Recordkeeping and Bias Testing Now Carry Legal Weight
The regulations also impose a four-year recordkeeping requirement covering ADS-related data, including the criteria a tool used to evaluate candidates, the scores or rankings it produced, and any bias testing performed on the system. Employers are not required to conduct bias audits, but the regulations make it relevant whether an employer tested its tools for discriminatory impact, and how thoroughly, in defending a discrimination claim.
An employer with no testing record and no documentation of how its ADS reached its outcomes will have a harder time defending a claim than one that can show it looked for bias and acted on what it found.
For employees and applicants, this recordkeeping requirement matters because it means the evidence of how an algorithm reached a decision relating to their employment should exist and be preserved. That evidence often becomes available only after a complaint is filed or a lawsuit is underway, which is one reason early legal advice matters in these cases.
What This Means If You Believe an Algorithm Discriminated Against You
If you were screened out of a job, denied a promotion, or terminated based in part on an automated scoring tool, and you believe race, age, disability, gender, religion, or another protected characteristic played a role in that outcome, you may have a claim under FEHA regardless of whether the discrimination was intentional. The law treats a biased algorithm the same way it treats a biased manager: if the practice produces a disparate impact on a protected group, it can be unlawful even without proof that anyone meant to discriminate.
Speak to an Employment Attorney at Keller Grover
If you suspect that an automated hiring or evaluation tool played a role in an unfair employment decision, an employment violations attorney at Keller Grover can help you understand your rights and evaluate whether the facts support a FEHA claim. Contact our office today to discuss what happened.