Job seekers suing Workday say its automated recruiting tools screened, ranked, or rejected applicants in ways that disadvantaged protected groups. The case, Mobley v. Workday, Inc., has survived important procedural challenges, but no court has found that Workday discriminated or that AI rejected each plaintiff. The dispute is about what the software did in practice—and whether a hiring-technology vendor can be held responsible when its tools influence access to jobs.
What is the lawsuit about?
Mobley v. Workday, Inc., case No. 3:23-cv-00770-RFL, was filed on February 21, 2023, in the U.S. District Court for the Northern District of California. Lead plaintiff Derek Mobley sued Workday, an enterprise human-resources software provider. The complaint invokes Title VII, the Americans with Disabilities Act (ADA), the Age Discrimination in Employment Act (ADEA), and California anti-discrimination law. The court’s case page identifies the proceedings.
The plaintiffs allege that Workday’s recruiting tools helped employers process applications and that the tools disproportionately harmed Black and Asian American applicants, women, people age 40 or older, and people with disabilities. These are allegations, not findings. The third amended complaint, filed March 27, 2026, describes a range of functions, including résumé parsing, assessments, predictive scoring, ranking, and automated dispositioning.
The Equal Employment Opportunity Commission (EEOC) filed an amicus brief on April 9, 2024, addressing whether federal employment-discrimination law can cover entities involved in screening or referring applicants for employers. The EEOC said it was not taking a position on whether the allegations against Workday were accurate. Its case page and brief explain its legal argument.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
What does “AI résumé screening” mean?
“AI” can describe different operations, and not every automated recruiting feature is generative AI. The complaint describes a broader screening process, not simply a chatbot reading résumés. Distinguishing the functions matters because extracting information is different from ranking or excluding a candidate.
- Résumé parsing: Extracting details such as skills, employers, dates, education, and job titles from an application.
- Rules-based screening: Checking answers or application data against requirements such as credentials, location, or work authorization.
- Matching and ranking: Comparing an applicant’s information with a role and assigning a suggested fit or priority.
- Assessments: Using tests or questionnaires to produce information or recommendations about candidates.
- Automated dispositioning: Applying configured rules to move an application to a rejected or inactive status.
- Generative AI: Producing text such as summaries, recommendations, or explanations.
Workday says its recruiting AI extracts relevant information from applications and résumés, compares it with job requirements, and may produce suggested grades such as “exceeds,” “meets,” or “does not meet some or all basic qualifications.” The company says those grades support recruiters rather than make the final hiring decision. That is Workday’s description of its products, not a court finding about how a particular employer configured or used them. See its recruiting AI explanation and recruitment privacy statement, effective June 3, 2026.
What does Workday say?
Workday’s public position is that its tools assist rather than replace human judgment, focus on qualifications and job requirements, and are not designed to use protected characteristics such as race, age, or disability. The company also says recruiters and hiring managers remain involved and that it conducts fairness testing and risk-management reviews. Its explanation of AI in hiring sets out that position.
Rank #2
The plaintiffs’ theory challenges the practical effect of the tools. A system need not directly receive a protected characteristic to raise a disparate-impact concern: other inputs—such as work history, education, location, language, or career gaps—may correlate with protected traits. That general possibility does not establish that it happened in this case. The central factual questions include what information a system used, how it was configured, how much weight recruiters gave its output, and whether a specific practice caused a legally significant disparity.
Why can a software company face an employment-discrimination claim?
Employers usually make the hiring decision, but the plaintiffs argue that Workday should also face responsibility because its software allegedly played a meaningful role in screening and referring applicants. The EEOC’s amicus brief argues that federal employment statutes can reach more than direct employers, including certain employment agencies and other entities that influence access to employment.
The legal theory may be more consequential when a vendor’s product does more than store records. There is a meaningful difference between supplying neutral infrastructure, letting an employer configure screening rules, recommending or ranking candidates, and automatically filtering applicants. The more a vendor’s system evaluates, refers, or excludes candidates, the stronger the argument—an inference, not a settled rule in this case—that the vendor influenced an employment decision. The court will still have to consider the applicable law and the evidence about the tools’ actual role.
What has the court decided—and what remains open?
On June 22, 2026, a federal judge allowed significant California-law and disability-related claims to continue, rejecting Workday’s effort to dispose of those claims at the pleading stage. The ruling also rejected Workday’s argument that California anti-discrimination law could not apply merely because applicants outside California sought jobs located elsewhere. Reuters reported on the ruling.
| Established at this stage | Not decided |
|---|---|
| Significant claims survived a challenge to whether they could proceed. | Whether Workday’s tools actually discriminated or caused unlawful disparities. |
| The software-vendor status alone did not dispose of the claims that continued. | Whether Workday, its customers, or both caused any particular applicant’s harm. |
| The court did not let the cited California-law argument end the claims that continued. | Whether the evidence will establish individual injury or support classwide or collective relief. |
| According to an American Bar Association analysis, a nationwide age-based collective action was conditionally certified in May 2025. | Whether plaintiffs will win, whether damages or an injunction are warranted, or whether the case will settle. |
Surviving a pleading-stage challenge means the claims were allowed to proceed; it is not a judgment that Workday is liable. It does not establish that every Workday customer used the same configuration, that every rejection involved AI, or that plaintiffs have proved intentional discrimination.
How could automated screening go wrong?
The following are general examples of possible failure modes, not findings about what happened to the plaintiffs:
- A résumé parser may miss relevant experience when it is described in unfamiliar terms or presented in an older format.
- A system may give less weight to equivalent skills gained through nontraditional work or foreign credentials.
- Career gaps associated with caregiving, illness, or disability may affect how a work history is interpreted.
- A knockout question may filter someone out because of an answer, a data-entry mistake, or a requirement that is not necessary for the role.
- A ranking system may reduce the chance of human review for applicants placed lower on a list.
- Historical hiring data may reproduce prior preferences if it is used to guide future recommendations.
A human-in-the-loop process does not automatically settle the question: review could be substantive or could rely heavily on an automated score. Conversely, a parser that only extracts text is not necessarily a decision-maker. The relevant issue is the system’s actual function and effect, not the label on the feature.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What can a job seeker do after a suspicious rejection?
A rejection alone does not show that AI was used or that discrimination occurred. If you have a concrete reason to question how an application was handled, preserve a record before details disappear.
- Save the application record. Keep the job posting, the résumé and cover letter you submitted, confirmation messages, screening questions and answers, rejection notices, dates, timestamps, and any assessment results or notices.
- Ask the employer about the process. You can ask whether automated tools were used, whether a person reviewed the application, and how to correct inaccurate application information. Whether the employer can or will provide details depends on the employer, vendor, jurisdiction, and applicable law.
- Read the employer’s privacy notice. Look for recruitment-data terms about profiling, automated decision-making, retention, correction, or appeals. A notice does not guarantee a right to inspect a proprietary model or obtain every internal record.
- Document a pattern, if there is one. Record repeated outcomes across comparable roles, relevant qualifications, later corrections to your application, or evidence that a person subsequently considered you qualified. A rapid rejection may prompt a question, but timing alone is not proof.
- Consider qualified help promptly. Depending on the facts and location, an employment lawyer, the EEOC, a state civil-rights agency, or a local legal-aid organization may be able to explain options. Legal deadlines can apply.
This case does not make every applicant who used Workday eligible to join a lawsuit. Whether someone can participate depends on the case’s orders and the person’s circumstances.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWhat employers and vendors should examine
For employers and technology providers, the dispute highlights a practical governance question: can they show how a screening tool affected candidates, and who was responsible for each step? Useful controls include:
- Testing selection outcomes for adverse impact and documenting the methods and results.
- Recording model inputs, outputs, configuration changes, and the reasons for human decisions.
- Making human review meaningful, with a route to question or override automated recommendations.
- Testing accessibility and whether assessments accommodate applicants with disabilities.
- Reviewing whether screening criteria are job-related, consistent with business necessity, and whether less discriminatory alternatives exist.
- Clarifying vendor and employer responsibilities in contracts, including access to audit records and incident response.
- Providing workable notice, correction, and appeal processes where required or appropriate.
What happens next?
The case still requires litigation over the evidence and legal questions, including causation, the significance of any disparity, the role of particular tools and customer configurations, and whether claims can proceed on a class or collective basis. The June 2026 ruling did not forecast a verdict or settlement. For job seekers and employers alike, the key unresolved issue is whether automated recruiting tools merely support decisions or materially shape who gets considered for work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
