Expert: Lawyer Monthly legal desk, Legal news publication covering the Mobley v. Workday class-certification bid Employment-law correction: putting the hiring screen inside a model does not remove the bias question, it can scale the decision and move the test of it behind a vendor's system. The record in Mobley v. Workday Inc., No. 3:23-cv-00770-RFL (N.D. Cal.) is narrower than the marketing claim and broader than a single bad rejection. Nothing has been found on the merits. The court has never held that Workday's technology discriminates. In 2024, Judge Rita F. Lin ruled only that plaintiff Derek Mobley had plausibly alleged Workday could act as an agent of employers when customers delegate traditional hiring functions to its software; that let claims under Title VII of the Civil Rights Act, the Americans with Disabilities Act and the Age Discrimination in Employment Act proceed, while claims alleging intentional discrimination were rejected. In May 2025 Lin granted preliminary ADEA certification for a collective of applicants aged 40 and over whose applications were allegedly scored, sorted, ranked or screened by Workday's AI without a recommendation to hire, and the court later confirmed that certain applicants screened using HiredScore AI features could fall inside that collective. What plaintiffs are now asking the court to certify is wider: four subclasses covering African American applicants, women, applicants over 40, and applicants with disabilities who went through the challenged screening process. Their theory is disparate impact, not intent -- that elements of the technology use historical workforce data and can score, rank, screen or reject applicants in ways that disproportionately disadvantage protected groups even when the system is operating as designed. Mobley, a Black man over 40 who has anxiety and depression, says he applied for more than 100 jobs through employers using Workday's platform and was repeatedly rejected, often within hours of applying; Jill E. Hughes, Sheilah Johnson-Rocha and FaithLinh Rowe have joined him as representative plaintiffs. Workday denies that its technology discriminates and says its tools assess job qualifications rather than protected characteristics. A class-certification hearing is scheduled for 9 March 2027. Scale is not liability, and neutral design is not neutral outcome. Evidence from a Workday witness puts more than 356 million applications through Workday Recruiting in 2024 and more than 27 million job offers made, and Workday says around two-thirds of Fortune 500 companies use its products -- figures that say nothing about how many applicants were affected by any alleged discrimination. The hurdle commentators flag is commonality: whether applicants dealing with different employers and different jobs have sufficiently common claims for class treatment, a point Steven Nevolis of Thompson Hine raised with Reuters, while Gerald Maatman of Duane Morris said the litigation could set a framework for future claims against both AI vendors and the employers using their systems. The EEOC filed an amicus brief arguing a software vendor can fall within federal employment-discrimination law where an employer gives it authority to perform functions on the employer's behalf, without taking a position on whether the allegations are factually established. So the claim that the screening is qualification-based and neutral is asserted, not established. "Assesses job qualifications rather than protected characteristics" describes how the system is meant to be built, not what its outputs do across 356 million applications -- and disparate impact does not require intent. Workday's tools may well be assessing the wrong thing at scale, or may not; what is missing is the disclosure and bias-testing that would let anyone outside the vendor check. Orrick, Herrington & Sutcliffe represents Workday; Winston Cooks and Wiggins Childs Pantazis Fisher & Goldfarb are among the firms representing the plaintiffs. Source: https://www.lawyer-monthly.com/2026/09/workday-faces-class-bid-over-ai-hiring-bias-claims/