What the Lawsuit Alleges
The lead plaintiff, Derek Mobley, is a Black man over the age of 40 who says he lives with depression and anxiety. According to the complaint, he has applied for more than 100 positions since 2017 through employers that use Workday’s platform and was rejected every time. Three additional plaintiffs have since joined the suit with their own rejection histories, according to Forbes.
One detail from the early filings has become a frequently cited example of how automated rejection can look from the applicant’s side. As the court noted in a 2024 order, Mobley allegedly received one rejection at 1:50 a.m., less than an hour after submitting his application, a timeline that suggested no human was reviewing his materials. The claims span federal law, including Title VII, the Age Discrimination in Employment Act (ADEA), and the Americans with Disabilities Act (ADA), as well as California’s Fair Employment and Housing Act (FEHA).
Why the Rulings Matter: The Vendor Is Not Just a Vendor
Workday’s core defense has been straightforward: it makes software, and its customers make the hiring decisions. The court has repeatedly declined to end the case on that basis.
In July 2024, the court allowed the plaintiffs to proceed on an “agent” theory of liability. As Staffing Industry Analysts reported at the time, the judge found that Workday could be treated as an employer because it carries out screening functions on behalf of the employers that use its software. The U.S. Equal Employment Opportunity Commission also filed amicus briefs supporting the plaintiffs’ position that Workday could be treated as an employer under the relevant statutes.
In May 2025, U.S. District Judge Rita Lin conditionally certified a nationwide collective under the ADEA, allowing older applicants who were screened through Workday’s system to opt in. The central question for that group is whether Workday’s recommendation system has a disparate impact on applicants over 40.
The most recent major ruling came on June 22, 2026. Judge Lin largely denied Workday’s motion to dismiss the plaintiffs’ Third Amended Complaint, as Reuters reported. Two parts of that decision stand out:
- California law can follow the software. The court rejected Workday’s argument that FEHA does not apply when its tools screen applicants outside California. Because the alleged discriminatory conduct stemmed from Workday’s California-based operations and decision-making, the judge found state law could potentially apply regardless of where the applicant or employer is located.
- Proxy discrimination is on the table. As Bloomberg Law reported, the court allowed a disability discrimination claim under the ADA to proceed based on allegations that Workday’s tools use proxy health indicators to screen out applicants.
Not every claim survived. The court dismissed a newly added race discrimination claim on behalf of Asian American applicants, finding it had not been properly added to the case or plausibly pled, along with certain claims tied to Workday’s role as a direct employer.
Workday continues to deny the allegations. In a statement reported by Reuters, the company said its technology considers only job qualifications rather than protected traits such as race, age, or disability, and that it rigorously tests its products through its Responsible AI program.
A separate discovery ruling in May 2026 may matter just as much to engineers and product teams. The court held that some of Workday’s AI bias-testing analysis could be shielded by attorney-client privilege because attorneys curated the underlying data. However, privilege did not extend to raw data or to reports filed with government agencies.
How AI Screening Tools Can Discriminate Without Meaning To
For a tech audience, the most important concept in the case is disparate impact. A screening model does not need to be programmed to consider age or disability to produce outcomes that fall harder on certain groups. Machine learning systems learn patterns from historical data, and if past hiring decisions reflected bias, a model trained on those decisions can reproduce it.
Proxy variables are the usual mechanism. Data points that look neutral on their face can correlate strongly with protected traits. Graduation years can track closely with age. Gaps in employment history can reflect medical leave, disability, or caregiving. Location data can correlate with race. A model that weights these signals heavily can screen out protected groups at higher rates even if no protected category ever appears in the training features.
That is why “we never told the model to look at age” is not a complete legal defense. Under disparate impact analysis, the question is the outcome, and whether the practice is job-related and consistent with business necessity.
California Is Writing the Rules in Real Time
The Mobley case is unfolding alongside some of the most aggressive state-level AI employment regulation in the country.
FEHA regulations on automated-decision systems. On October 1, 2025, new regulations from the California Civil Rights Council took effect, clarifying that employers cannot use automated-decision systems in ways that discriminate under FEHA. The Civil Rights Department’s notice of approval lays out the framework. Among the key provisions, employers must retain automated-decision system data for at least four years, evidence of anti-bias testing (or the lack of it) is relevant to defending a claim, and the definition of an employer’s “agent” is broad enough that the owner or licensor of a screening tool may qualify.
The No Robo Bosses Act (SB 947). On September 30, 2026, Governor Gavin Newsom signed SB 947. The governor’s office announced the signing as part of a package of worker protections, and the full text of SB 947 is available from the California Legislature. Beginning July 1, 2027, employers will be barred from relying solely on an automated decision system to make disciplinary or termination decisions. Where automated output is the primary basis for such a decision, human corroboration is required.
Surveillance and layoff transparency. The same package included AB 1883, which restricts AI workplace surveillance tools that collect employees’ neural data or predict their emotional states, and SB 951, which addresses notice and reporting when AI or automation contributes to mass layoffs. Tech Policy Press has a rundown of the full slate of AI bills from this year’s session.
What This Means for Tech Companies
For companies that build hiring software, the message from the court so far is that “we only make the tool” may not hold up if the tool is effectively making the decision. For companies that buy that software, outsourcing the screen does not outsource the legal risk. Practical steps commonly recommended for compliance include:
- Inventorying every AI or algorithmic tool used in recruiting, hiring, promotion, discipline, and termination.
- Running and documenting anti-bias testing before deployment and at regular intervals afterward.
- Reviewing vendor contracts for transparency, audit rights, and responsibility for discriminatory outcomes.
- Building meaningful human review into decisions, rather than rubber-stamping model outputs.
- Retaining automated-decision data for at least the four years California’s regulations require.
What Job Seekers and Employees Can Do
Most applicants never learn why they were rejected, and many never learn that software made the first cut. If you suspect an automated system screened you out unfairly, a few habits can help preserve your options:
- Keep records. Save job postings, confirmation emails, and rejection notices, including the time stamps. A rejection that arrives minutes after you apply can be meaningful evidence.
- Track patterns. Note which employers and which application platforms you used. Repeated rejections across employers using the same system can matter.
- Know the deadlines. In California, workers generally have three years to file a FEHA complaint with the California Civil Rights Department. The deadline to file a charge with the U.S. Equal Employment Opportunity Commission is much shorter, typically 300 days in California.
- Talk to an employment lawyer. AI discrimination claims involve unsettled law and technical evidence. Bay Area workers who believe they were passed over, disciplined, or fired because of age, race, disability, sex, or another protected trait can get guidance from experienced San Francisco employment and civil rights attorneys such as Lawless Lawless & McGrath, who represent employees in discrimination, wrongful termination, and retaliation matters.
What Comes Next
Workday has not been found liable, and the case still faces discovery fights over how its screening features actually work, along with class and collective certification questions down the road. Whatever the outcome, Mobley v. Workday has already established that AI hiring tools will be judged by the same anti-discrimination standards as human recruiters. With California’s new rules phasing in through 2027, the companies building and buying these systems in the state that invented most of them are now on notice.
Disclaimer: This article is provided for general informational purposes only and does not constitute legal advice. Reading this article does not create an attorney-client relationship. Laws and court rulings change, and the facts of every situation are different. If you believe you have experienced employment discrimination, consult a licensed attorney in your state about your specific circumstances. The allegations described in this article have not been proven in court.