One particular aspect of the Mobley v. Workday lawsuit has stuck with employment attorneys who have studied the case. Derek Mobley, the plaintiff, submitted an application for a position via the employment platform’s Workday-powered system. Less than an hour after submitting the application, Mobley received his rejection email at 1:50 a.m. In any employment procedure that involves human evaluation, it is nearly impossible to explain the timestamp.
Practically speaking, it is the fingerprint of an algorithm that makes a screening decision on a timeline that no human would adhere to and at a speed that no hiring manager could equal. A qualified applicant over 40 who received a rejection email at 1:50 a.m. is now at the center of what appears to be the first significant countrywide collective action against algorithmic discrimination in American employment law.
| Category | Detail |
|---|---|
| Mobley v. Workday, Inc. | Case No. 3:23-cv-00770 (N.D. California); Judge Rita Lin; preliminary collective certification granted May 16, 2025; notice plan approved December 2, 2025; plaintiff Derek Mobley alleged rejection from over 100 positions via Workday’s AI screening, including a rejection received at 1:50 a.m. within an hour of application |
| Legal Basis | Age Discrimination in Employment Act (ADEA) — protects workers 40 and older; disparate impact theory allows claims even without proof of intentional discrimination; plaintiff joined by four opt-in co-plaintiffs; preliminary class includes all applicants aged 40+ rejected through Workday’s platform since September 24, 2020 |
| Vendor Liability Doctrine | Court’s July 12, 2024 ruling rejected “employment agency” theory but allowed “agent” theory to proceed — establishing that AI vendors can be directly liable alongside employers when their product participates meaningfully in consequential hiring decisions |
| Harper v. Sirius XM Radio | Parallel 2025 case — plaintiff Arshon Harper alleges his 149 job applications were rejected by Sirius XM’s AI hiring tool in violation of Title VII; claims his qualifications met or exceeded the stated requirements for positions denied |
| EEOC Position | Equal Employment Opportunity Commission filed amicus brief in April 2024 supporting the position that algorithmic hiring tools can violate anti-discrimination laws without explicit intent and that vendors can be held accountable alongside employers |
| HiredScore Inclusion | Court held on July 7, 2025 that the preliminary collective includes applicants screened using Workday’s HiredScore AI features, rejecting Workday’s arguments that HiredScore was acquired separately and used different algorithms |
| Beyond Hiring — Other Domains | Active legal risk extends to housing (tenant-screening algorithms), lending (credit scoring models), and financial services — all areas where algorithmic systems make consequential decisions using data that may reflect historical disparities |
| Further Reference | Legal analysis at the University of Miami Law Review and the Seyfarth Shaw AI litigation tracker |
For about two years, civil rights and employment lawyers have been alerting people to the impending flood. In an amicus brief filed in April 2024, the Equal Employment Opportunity Commission stated unequivocally that algorithmic hiring tools may violate anti-discrimination laws even in the absence of evidence of deliberate bias, and that both employers and the vendors who create and market these tools may be held liable. In 2024 and 2025, Judge Rita Lin of the Northern District of California took up that thread and issued a number of decisions that together changed the legal landscape.
Workday’s move to dismiss was denied by the court on July 12, 2024, and claims against the AI provider itself were permitted to proceed under the “agent” theory of responsibility. Mobley and four opt-in plaintiffs are now able to prosecute the case on behalf of all Workday candidates 40 years of age and over who have been rejected employment recommendations since September 24, 2020, thanks to Judge Lin’s preliminary certification of a countrywide collective action under the ADEA on May 16, 2025.
The scale is important. Thousands of employers throughout the world utilize Workday. Some plaintiffs’ lawyers claim that the preliminary aggregate may include hundreds of millions of applicant data over the pertinent time frame. A notification strategy that is now being used to notify potential collective members of their right to opt in was granted by the court on December 2, 2025.
Additionally, a July 7, 2025 court rejected Workday’s claim that HiredScore, which was bought later, employed distinct algorithms and should be treated separately, holding that the collective comprises applicants screened by Workday’s HiredScore AI features. There is yet more to be discovered. A verdict on the merits may be reached in 2026, although settlement prior to trial is still feasible and, in cases of this magnitude, maybe likely.
There are other ongoing cases involving algorithmic prejudice besides Mobley. Arshon Harper sued Sirius XM Radio in 2025, claiming that despite his qualifications meeting or surpassing the specified requirements, the company’s AI hiring tool had rejected 149 of his job applications. He attributes this pattern to racial discrimination under Title VII of the Civil Rights Act.
Numerous counties have filed lawsuits contesting tenant-screening algorithms, claiming that these systems disproportionately reject applicants of color based on eviction records and criminal history information, which itself reflect previous racial inequities. Federal courts are also considering lending discrimination lawsuits that use comparable disparate impact doctrines. The trend is so obvious that law firms that represent big employers have started to issue comprehensive compliance guidelines, including as independent bias audits, significant human oversight, and thorough vendor due diligence.
There’s a feeling, watching these cases develop, that employment and civil rights law is entering a phase where its core concepts — disparate impact, constructive notice, vendor agency — are being applied to a category of decision-making the drafters of the original statutes could not have imagined. Title VII was enacted in 1964.
The ADEA followed in 1967. Neither anticipated automated systems that could screen hundreds of thousands of applicants per day, making decisions in milliseconds using data patterns that reflect decades of pre-existing discrimination in hiring and employment. What the courts are doing, through cases like Mobley v. Workday, is stretching existing civil rights doctrine to cover algorithmic decision-making.
This span might turn out to be legally durable. It’s also possible that specific legislation — more targeted to AI’s particular characteristics — will eventually supplement or replace the current patchwork. The fact that algorithmic choices fall under the purview of anti-discrimination law is no longer up for debate. The 1:50 a.m. rejection email turned out to be a piece of evidence that moved the law.
