Case Summary

1. Collective Certified: District Judge Rita Lin approved preliminary certification for a nationwide group of applicants aged 40‑plus who say Workday’s résumé‑ranking engine shut them out of jobs.
2. Unified Policy in the Crosshairs: The order finds Workday’s Candidate Skills Match and Assessment Connector tools form a single, AI‑driven policy whose disparate impact can be proven with common evidence.
3. Ripple Risk for HR Tech: The ruling rejects Workday’s “too big to notify” argument and signals that vendors—not just employers—may face sweeping liability if algorithms skew against protected groups.

By Samuel Lopez – USA Herald

SAN JOSE, CA - When Derek Mobley clicked “submit” on yet another Workday‑powered application, he had no idea he was stepping into an AI minefield. Five years and hundreds of automated rejections later, his lawsuit is poised to redefine the legal landscape for every résumé‑screening algorithm in America.

On May 16, 2025, Judge Rita Lin issued a 33‑page order that grants preliminary collective certification to “[a]ll individuals aged 40 and over who, from September 24, 2020 to the present, applied for jobs using Workday and were denied employment recommendations.”  The decision gives Mobley and four co‑plaintiffs the green light to alert potentially millions of similarly situated job‑seekers.

Inside the Order: Five Takeaways Employers Can’t Ignore

1. A Single, AI‑Driven Policy Is Enough

Workday insisted its customers, not its software, make hiring decisions. Judge Lin disagreed, pointing to discovery excerpts showing that Candidate Skills Match (CSM) automatically assigns “strong,” “good,” “fair,” or “low” scores, while Workday Assessment Connector (WAC) can down‑rank protected‑class applicants when employer behavior trains it to do so. That unified scoring framework—“the digital gatekeeper,” as the court puts it—creates a common question for every applicant over 40.

2. Sliding‑Scale Burdens Rejected

Workday urged a tougher evidentiary standard because some discovery had occurred. The court refused, sticking to the Ninth Circuit’s two‑step collective action test, where stage one is “loosely akin to a plausibility standard” and stage two resembles summary judgment NYC.gov. Translation: Plaintiffs do not need full statistics now; credible allegations plus a coherent theory suffice.

3. Size Doesn’t Scare the Court

Workday warned that notifying every rejected applicant could swamp the docket with “hundreds of millions” of claims. Judge Lin answered bluntly: allegedly widespread discrimination is not a reason to gag notice.” If Workday’s own records show 1.1 billion rejections, so be it—the company must still cooperate on a notice plan by May 28, 2025, ahead of a June 4 case‑management conference EEOC.

4. Data‑Access Hurdles Fall Away

Workday argued privacy clauses bar it from mining applicant data. The court cited a “court‑order exception” in those contracts and pointed to proxies such as graduation year to identify age‑40+ candidates Artificial Intelligence Act. Employers using Workday should brace for subpoenas seeking the same fields.

5. Disparate Impact, Not Intent, Runs the Show

Unlike disparate‑treatment claims, plaintiffs do not need to prove intentional bias or personal qualifications at this stage. The injury is being forced through a filter that statistically undervalues older workers. That subtle shift aligns with Supreme Court precedent on biased testing and opens the door for statistical experts to attack AI models themselves.

Wider Implications: Algorithms Under the Microscope

Vendors on the Hook

For decades, discrimination lawsuits targeted employers. By framing Workday as a joint decision‑maker, Judge Lin places HR‑tech providers squarely in the liability zone. Any vendor whose algorithm scores or ranks candidates—Oracle, SAP SuccessFactors, iCIMS—must now validate and audit models or risk collective actions.

Regulatory Tailwinds

• EEOC Guidance:In 2023 the EEOC warned employers that algorithmic tools creating disparate impact violate Title VII, even if supplied by a third party.
• NYC Local Law 144:Since July 5, 2023, New York City bars employers from using Automated Employment Decision Tools without a public bias audit.
• EU AI Act:Europe’s new regulation classifies CV‑scanning tools as “high risk,” imposing transparency, human oversight, and record‑keeping requirements.

Lessons from Amazon’s Ghost Algorithm

Amazon famously scrapped an internal recruiting engine after it downgraded résumés containing the word “women’s” or graduates of all‑female colleges. The parallels to Workday underscore how training data baked in yesterday’s bias can sabotage tomorrow’s talent pipeline.

What This Means for Job‑Seekers Over 40

• Watch Your Inbox:If you applied through Workday since 2020 and got a quick rejection, you may soon receive court‑approved notice inviting you to opt in.
• Save the Evidence:Keep timestamped rejection emails, screenshots of “low match” dashboards, and notes about interview denials. Such artifacts bolster disparate‑impact statistics.
• Potential Remedies:Opt‑in plaintiffs could secure injunctive relief—forcing algorithmic audits—and monetary damages tied to lost wages.

Compliance Playbook for Employers

1. Conduct Independent Bias Auditsannually and publish summaries.
2. Enable Explainability:Provide candidates with plain‑language reasons for rejections.
3. Add Human Review:Offer an appeal path where human recruiters override low scores.
4. Contractual Shields:Clarify indemnity and data‑sharing terms with vendors, anticipating subpoenas.
5. Document Validation:Maintain contemporaneous studies showing the algorithm predicts job success without adverse impact.

Ignoring these steps courts disaster; Judge Lin’s order all but invites plaintiffs’ firms to test every major hiring platform.

All allegations remain unproven. Workday contends the claims are meritless and will seek decertification once discovery is complete.

Explore More
🔗 Follow us on X @RealUSAHerald