An African-American applicant rejected 100 times. A deaf candidate scored by facial analysis software. Age cutoffs coded directly into screening algorithms. These are not hypotheticals — they are active federal cases exposing how AI hiring tools discriminate at scale.
The Numbers That Should Alarm You
In 2025, 99% of Fortune 500 companies used some form of automated system to filter job applications. Roughly 83% of all companies relied on AI to screen resumes before a human ever saw them. And 76% planned to deploy AI for conducting interview questions directly. These are not projections. These are the numbers from industry surveys published over the past year.
Here is the statistic that makes those figures dangerous: 67% of the same companies acknowledged that their AI hiring tools likely contain bias. They deployed them anyway.
That gap between awareness and action isn’t an oversight. It’s a calculation. AI screening tools save companies an estimated 31% in hiring time and claim a 50% improvement in “quality of hire” metrics. When efficiency conflicts with fairness, efficiency wins — until someone sues.
Someone has been suing. In fact, a growing wave of federal lawsuits is pulling back the curtain on exactly how these systems discriminate, who they hurt, and why the companies using them looked the other way.
Case File: Derek Mobley vs. Workday
Derek Mobley applied for over 100 jobs through Workday’s platform between 2017 and 2023. He is African-American, over 40, and has anxiety and depression. He received zero offers.
In May 2025, the U.S. District Court for the Northern District of California certified a collective action in Mobley v. Workday, Inc. — a landmark decision that treats Workday not merely as a software vendor but as a potential agent of employment discrimination. The court ruled that Mobley’s claims under Title VII of the Civil Rights Act, the Age Discrimination in Employment Act, and the Americans with Disabilities Act could proceed as a class action.
The allegations are specific and technical. Workday’s AI-powered screening tools, marketed as “smart” hiring solutions, used historical hiring data to evaluate candidates. The problem with historical data, of course, is that it encodes historical bias. If past hiring decisions disproportionately favored younger, white, non-disabled candidates, then an algorithm trained on those decisions will replicate and amplify those preferences — at machine speed and machine scale.
Mobley’s legal team argues that Workday’s system effectively launders human prejudice through mathematical optimization. The algorithm doesn’t see race or age explicitly. It sees proxies — graduation dates, zip codes, university names, employment gaps — that correlate with protected characteristics closely enough to produce discriminatory outcomes without ever using a prohibited variable.
Scale of impact: Workday processes hiring workflows for more than 10,000 organizations worldwide, including over 50% of the Fortune 500. If its screening algorithm carries systemic bias, the number of affected applicants could reach into the millions. The opt-in deadline for the class action is March 7, 2026.
More Cases, Same Pattern
Mobley’s case is the most prominent, but it is far from the only one. The pattern repeating across federal courts is remarkably consistent: an AI tool is deployed, it produces discriminatory outcomes, and the company using it claims it didn’t know — or didn’t intend — for that to happen.
Harper v. Sirius XM Radio (2025). Filed in the Eastern District of Michigan on August 4, 2025, this case alleges that Sirius XM’s AI-powered hiring system discriminated against a Black applicant by relying on historical hiring data that perpetuated past racial biases. The lawsuit is significant because it targets the employer directly, not the AI vendor, arguing that companies cannot outsource their civil rights obligations to algorithms.
ACLU v. HireVue/Intuit (2025). In March 2025, the ACLU of Colorado filed an EEOC complaint against Intuit and its AI vendor HireVue. The complaint centers on HireVue’s automated video interview system, which records candidates, analyzes their speech and facial expressions, and produces numerical scores. The ACLU alleged the system was inaccessible to deaf applicants and performed measurably worse when evaluating non-white candidates, particularly those who spoke dialects like Native American English or African American Vernacular English.
ACLU v. Aon (2025). Another ACLU challenge targets three of Aon’s hiring assessment tools: ADEPT-15 (a personality test), vidAssess-AI (a video analysis system), and gridChallenge (a gamified cognitive test). The complaint alleges these tools discriminate against people with disabilities and certain racial groups, and that Aon’s marketing of them as “bias-free” constitutes deceptive trade practices.
iTutorGroup Settlement (2023). Perhaps the most straightforward case: the online tutoring company programmed its AI recruitment software to automatically reject female applicants over 55 and male applicants over 60. No subtlety, no proxy variables — raw age cutoffs in the code. The EEOC brought the case, and iTutorGroup paid $365,000 to settle after the AI had rejected over 200 qualified applicants based solely on their age.
AI Hiring Bias: Case Tracker
Why Bias Audits Aren’t Enough
New York City’s Local Law 144, enforced since July 2023, was supposed to be the solution. It requires companies using automated employment decision tools to conduct annual bias audits and publish the results. On paper, it sounded like accountability.
In practice, it has been a lesson in regulatory theater. NYC’s Department of Consumer and Worker Protection has struggled to identify companies that are even subject to the law, let alone enforce it. Many employers simply don’t disclose that they’re using AI in hiring. Others publish audit results so vague they’re meaningless — confirming that the tool was tested without revealing what the tests actually found.
The deeper structural problem is that bias audits examine outcomes, not mechanisms. An audit might reveal that a tool approves white candidates at a rate 15% higher than Black candidates. It will not explain why, because the “why” is encoded in millions of weighted connections within a neural network that no auditor can fully interpret. The tool passes the audit by adjusting its outputs to fall within acceptable disparate impact thresholds, while the underlying model — trained on the same biased data — remains unchanged.
Illinois took a different approach with House Bill 3773, effective January 1, 2026. Rather than mandating audits, Illinois requires employers to notify applicants whenever AI is used in hiring, promotion, discipline, or termination decisions. The law also prohibits discriminatory outcomes regardless of whether the employer intended them — a strict liability standard that shifts the burden from proving intent to proving impact.
| Law / Action | Jurisdiction | Approach | Limitation |
|---|---|---|---|
| NYC Local Law 144 | New York City | Annual bias audits, public disclosure | Weak enforcement, vague results |
| Illinois HB 3773 | Illinois | Mandatory notification, strict liability | No audit requirement |
| EU AI Act | European Union | High-risk classification, fines up to 7% revenue | Full enforcement begins Aug 2026 |
| EEOC Guidance | Federal (U.S.) | Existing civil rights law applies to AI tools | Case-by-case litigation |
| Illinois BIPA | Illinois | Written consent for biometric collection | Narrow scope (biometrics only) |
How to Protect Yourself as a Job Applicant
If you’re applying for jobs in 2026, you’re almost certainly being evaluated by AI at some point in the process. You may not be told. You may not notice. But there are steps that shift some power back in your direction.
Ask directly. In jurisdictions like Illinois and New York City, employers are legally required to disclose AI use in hiring. Even where disclosure isn’t mandated, asking “Is AI or automated software used to evaluate my application?” creates a documented record. If the answer is yes, ask what data points the system evaluates and whether a human reviews its recommendations.
Request your file. Under the EEOC’s existing framework and under emerging state laws, you may have the right to request the data an employer or its AI vendor holds about you. This includes any scores, rankings, or classifications the AI assigned. If you were rejected, this information can reveal whether the decision was based on legitimate qualifications or on proxy variables correlated with protected characteristics.
Document everything. If you suspect AI-driven discrimination, keep records of every application, every automated response, and every rejection. Note the timeline and any patterns. The Mobley v. Workday case succeeded in part because the plaintiff documented over 100 rejections — a volume that made individual bad luck implausible and systemic bias probable.
Know your rights by state. AI hiring discrimination law varies dramatically by jurisdiction. Illinois, New York City, Maryland, and Colorado have specific AI-in-hiring regulations. California’s CCPA gives you rights over automated decision-making. Federal civil rights laws (Title VII, ADA, ADEA) apply everywhere, though enforcement requires litigation. If you believe you’ve been discriminated against, contact the EEOC or a state civil rights agency.
Consider the EEOC. The Equal Employment Opportunity Commission has made AI hiring bias an enforcement priority. Filing an EEOC charge is free, doesn’t require a lawyer, and creates a federal record. Even if your individual case doesn’t result in action, the pattern data from multiple complaints helps the EEOC identify systemic problems and bring larger enforcement actions.
Frequently Asked Questions
It depends on your jurisdiction. In New York City (Local Law 144) and Illinois (HB 3773, effective January 2026), employers must disclose AI use in hiring. Maryland requires consent before using facial recognition in interviews. However, most U.S. states have no specific AI hiring disclosure law yet. Federal civil rights protections still apply — if an AI tool produces discriminatory outcomes, the employer can be liable under Title VII regardless of whether they disclosed the AI’s involvement.
Yes, but the path matters. You generally cannot sue for being screened by AI alone — you need evidence that the screening produced a discriminatory outcome based on a protected characteristic (race, gender, age, disability). The strongest approach is filing an EEOC charge first, which triggers an investigation and preserves your right to sue. Class actions like Mobley v. Workday are powerful because they demonstrate patterns across many applicants. Individual claims are harder but not impossible, especially in states with strict liability standards like Illinois.
The technology is improving, but the fundamental problem persists. AI tools trained on historical hiring data will reflect historical biases unless developers take deliberate, ongoing steps to counteract them. Some vendors now use adversarial debiasing techniques and synthetic data augmentation, which help. But 67% of companies using AI screening still acknowledge bias concerns, and the growing number of federal lawsuits suggests the problem is far from solved. The most honest assessment: the tools are getting more sophisticated at hiding bias, not necessarily at eliminating it.