Photo by Vitaly Gariev on Unsplash
- As of May 16, 2025, Federal Judge Rita Lin certified Mobley v. Workday, Inc. as a collective action — potentially the largest AI hiring discrimination case in U.S. legal history.
- Workday's own court filings disclosed that its AI screening systems rejected 1.1 billion job applications during the class period beginning September 24, 2020.
- The court found that Workday — not only the employers using its software — can face direct liability as an "agent" under the Age Discrimination in Employment Act (ADEA).
- Job seekers aged 40 or older who were screened through Workday's platform since September 2020 had until March 7, 2026 to opt into the collective action at workdaycase.com.
What Happened
1.1 billion. That number — disclosed by Workday in its own court filings — represents the volume of job applications its AI screening tools rejected during roughly five years starting September 24, 2020. As of June 17, 2026, that figure sits at the center of what may become the most consequential employment discrimination case of the AI era, according to reporting aggregated by Google News and corroborated across multiple employment law outlets.
The case is Mobley v. Workday, Inc. (Case No. 3:23-cv-00770, N.D. Cal.), brought by Derek Mobley — a Black man over the age of 40 with disabilities — who alleged that he was rejected for more than 100 jobs processed through Workday's platform, often within minutes of applying or in the middle of the night, with no apparent human review at any stage. Workday's AI screening product, HiredScore, scores and ranks applicants against job criteria before a human recruiter ever sees the pile.
On May 16, 2025, Federal Judge Rita Lin granted conditional certification, allowing the lawsuit to proceed as a collective action. A March 6, 2026 ruling then cut in two directions: certain California state law claims and individual disability allegations were dismissed, while the federal ADEA age discrimination claims advanced. An amended complaint followed on March 30, 2026. Eligible workers — applicants aged 40 or older screened through Workday since September 2020 — had until March 7, 2026 to submit opt-in forms at workdaycase.com.
The Algorithm as Defendant — How the Court Got Here
The ADEA, enacted in 1967, prohibits employers and their agents from discriminating against workers aged 40 and over. Courts have long applied that agent designation to staffing firms and background-check vendors. Judge Lin extended the same logic to an AI software company — holding that Workday's system, which actively scores, sorts, ranks, and screens applicants, functions as an agent of the hiring employer, not merely a passive database query.
That distinction reshapes the entire legal technology vendor landscape. Workday's stated defense is that hiring decisions ultimately rest with its employer-clients and that HiredScore "simply compares candidate qualifications to client-defined job requirements" without identifying or acting on protected characteristics. The court declined to treat that argument as dispositive at the certification stage. Judge Lin wrote that "allegedly widespread discrimination is not a basis for denying notice" — meaning that the sheer scale of rejections could not itself be used to block class members from receiving notice of the lawsuit.
Legal analysts tracking the case across multiple employment law reviews have noted that "the court's decision to let the case proceed sends a clear message that employers and vendors can be held responsible when their AI systems cause discriminatory outcomes, even if it's unintentional." That last word carries legal weight. The case rests partly on disparate impact theory: the principle that a facially neutral policy can still violate civil rights statutes if it produces outcomes that disproportionately harm a protected class. The ADEA has not been amended to remove that theory.
The regulatory environment, however, is unsettled. The EEOC removed all of its AI-related employment guidance from its website on January 27, 2025, reversing May 2023 guidance that had supported claims like Mobley's. And on April 23, 2025, President Trump signed an executive order directing federal agencies to walk back disparate impact enforcement across civil rights law broadly. Courts, though, apply existing statutes as written — and what those statutes say has not changed. As SmartCareer AI observed in its coverage of what workers can actually do in a difficult job market, navigating AI-driven hiring systems has become a front-line employment concern entirely separate from regulatory debates.
The Numbers Behind the Case
Chart: HiredScore's claimed efficiency gains per Workday marketplace documentation — the same speed metrics that, at scale, translate to 1.1 billion automated rejections. Source: Workday marketplace documentation, as of June 17, 2026.
HiredScore claims to reduce screening time by 55% on average and hiring manager review time by 35%, according to Workday marketplace documentation. Those efficiency numbers are, in one sense, the product's entire value proposition. They are also the lawsuit's central problem. A human recruiter with an unconscious preference against candidates over 40 might affect dozens of applications in a quarter. An algorithm with the same bias — or one trained on historical hiring data that encodes past discrimination — affects hundreds of millions of applicants before anyone notices the pattern.
The HR software market is projected to exceed $30 billion by 2026, with AI screening tools processing billions of applications annually. At that scale, even a small systematic error compounds into mass harm. Colorado's AI Act (SB 24-205), which would require impact assessments for "high-risk" AI systems including hiring tools, was delayed from its February 1, 2026 effective date to June 30, 2026, illustrating how far behind regulation remains relative to deployment. Amazon discontinued its own AI recruiting tool back in 2017 after discovering it systematically downgraded resumes from women's colleges — a cautionary precedent that foreshadowed exactly the structural dynamics now at issue in Mobley. A separate Workday-related matter, a $15 million settlement in Oregon for payroll system errors affecting state employees from December 2022 through June 2025, adds context about the company's broader operational risk profile, though that dispute involved different systems entirely.
Where You Stand If You've Applied for Jobs Online
The collective action opt-in deadline of March 7, 2026 has already passed as of June 17, 2026. If you missed it, joining the Mobley collective action as currently constituted is no longer an option — though individual EEOC charges and parallel state-law claims may still be available depending on your circumstances and jurisdiction. That is a question for an employment attorney, not a blog post.
What this case establishes regardless of its final outcome is a framework for evaluating algorithmic hiring discrimination going forward. If you are 40 or older and have experienced a pattern of near-instant rejections from employers using AI-mediated applicant tracking systems — particularly Workday's platform — the type of evidence that mattered in Mobley is worth collecting now: rejection timestamps, positions applied for, and whether any human follow-up ever occurred. A rejection at 2:47 a.m. with no subsequent contact is a different kind of data point than a rejection letter from a named recruiter.
Employers relying on AI legal tools and automated screening platforms should take the agent liability theory seriously. The court's reasoning does not limit exposure to Workday — it applies the ADEA's agent framework to any vendor whose system actively ranks, scores, or filters candidates. If your HR department's screening process runs through software that recommends or de-prioritizes applicants, the question "could this produce disparate outcomes by age, race, or disability?" is now a compliance question with direct litigation exposure behind it, not a theoretical ethics concern.
Three Steps If You Think AI Screened You Out
Document every automated rejection you receive: the platform visible in the application portal, the exact time of day the rejection arrived, and how quickly it followed your submission. Rejections arriving within minutes or overnight with no subsequent human contact are the pattern Judge Lin's court found probative in the Mobley certification analysis. A simple spreadsheet tracking date, employer, position, time submitted, and time rejected is a starting point for any future legal claim and costs nothing to maintain.
Under California's CPRA and similar statutes in other states, you can formally request the personal data a company holds about you — including scoring or ranking data generated by automated systems. If a company used Workday's platform, that request may surface whether and how HiredScore scored your application. The statute reads differently by jurisdiction, so verify what rights apply in your state before submitting a request. Some states require responses within 30 to 45 days.
A court would likely look at whether the rejection itself — or a pattern of rejections — starts the ADEA's limitations clock. Under federal law, ADEA charges typically must be filed with the EEOC within 180 or 300 days of the discriminatory act, depending on the state. State law timelines vary further. A free or low-cost consultation with an employment attorney is not a commitment to litigation — it is a way to find out whether you have a deadline approaching without knowing it.
Frequently Asked Questions
Who qualifies for the Workday AI age discrimination lawsuit?
The collective action in Mobley v. Workday was open to job applicants aged 40 or older whose applications were processed and screened by Workday's AI systems between September 24, 2020 and the filing period. The court-ordered opt-in deadline was March 7, 2026, submitted via workdaycase.com. As of June 17, 2026, that opt-in window has closed for the current collective action phase. Separate individual EEOC charges or state-law claims may still be viable depending on your specific situation and jurisdiction — an employment attorney can evaluate whether a separate filing makes sense.
Can I sue for algorithmic hiring discrimination even if the employer says the AI just made a recommendation?
That is exactly the argument Mobley v. Workday is testing. Judge Lin's certification ruling held that Workday's system — which actively scores, sorts, and ranks applicants — functions as an "agent" under the ADEA, not merely a neutral tool. The employer's argument that a human made the "final" decision does not necessarily insulate either the employer or the software vendor from liability if the AI's recommendations effectively drove the outcome. Whether that theory survives trial or produces a settlement will shape how future cases are handled. For your own situation, the presence or absence of any human review before rejection is a key factual question.
Is Workday liable for AI discrimination in hiring, or only the companies that use it?
As of June 17, 2026, Judge Rita Lin's ruling in the Northern District of California holds that Workday can face direct liability as an agent under federal law — not merely as a third-party vendor shielded by its employer-clients. Workday contests this, arguing that hiring decisions rest with client employers and that its technology compares qualifications to client-defined criteria without targeting protected characteristics. The court found that argument insufficient to block collective action certification. Final liability will be determined at trial or in any settlement, but the ruling itself is precedent-setting for how legal technology vendors structure their contracts and compliance obligations.
In my read, the most consequential aspect of Mobley v. Workday is not any individual plaintiff's claims — it is the vendor liability framework the court articulated. If this case produces a major settlement or survives to verdict, every HR software company selling AI screening capabilities will face pressure to demonstrate bias auditing, renegotiate indemnification terms with employer-clients, and reconsider whether efficiency metrics in their own marketing materials create evidentiary exposure. The era of AI vendors treating algorithmic bias liability as the customer's compliance problem alone is likely ending — and that shift will ripple through every enterprise contract in the space.
Explore Our Network
Disclaimer: This article is for informational purposes only and does not constitute legal advice. Individual circumstances vary; consult a licensed employment attorney in your jurisdiction before taking any legal action. Research based on publicly available sources current as of June 17, 2026.
