Did AI Really Break Hiring? The Truth Behind the Workday Lawsuit
Derek Mobley applied to over a hundred jobs. He was rejected from every single one. The rejection emails often came in the middle of the night. He suspected no human was making these decisions. He was right. Mobley is Black, over 40, and lives with anxiety and depression. Instead of suing the companies that rejected him, he sued Workday, the software company behind the rejections. In June 2024, a federal judge ruled that Workday must answer for this pattern of discrimination. This case shows that AI didn’t break hiring. It simply scaled the bias we already chose.
How AI Amplifies Old Bias
Every hiring process has always been biased. Humans run it, and humans are biased. This isn't about a few bad apples. It's about everyday, structural bias that doesn't require cruelty. Most organizations don't measure bias in hiring because fixing it is expensive and uncomfortable. Now, AI has taken these hidden biases and turned them into a consistent, automated rule.
What Derek Mobley’s Case Reveals
Workday argued that it just makes the software; employers make the decisions. Judge Rita Lin rejected that claim. She ruled that Workday builds, trains, and runs these tools from its California headquarters. So it can't avoid California’s anti-discrimination laws. The disability claim is especially important. Mobley’s lawyers argue the software uses proxies like gaps in employment history to screen out people with medical leave. No HR manual says “reject cancer survivors.” But AI can learn that pattern from old data and apply it silently at scale.
Why We Should Have Seen This Coming
AI didn't invent new discrimination. It learned from decades of real hiring data. The algorithm mirrors the reality where some people advanced and others stalled. It simply reproduced that reality faster and more consistently. Today, over 80% of U.S. employers use AI hiring tools. The Fortune 500 is no exception. Yet most companies never checked what biases these tools were copying.
The Real Problem: We Ignored Bias for Decades
Women run only 11% of Fortune 500 companies. Board representation has stalled. People of color hold fewer than one in seven top corporate jobs. These aren't AI's mistakes. They are the result of thousands of “neutral” decisions made by well-meaning people. AI just made the pattern visible. A decade ago, a man like Mobley could be rejected by a hundred employers too—but slowly, by different people in different offices. No one saw the pattern. Now one system applies one logic instantly, making discrimination trackable.
What HR Leaders Must Do Now
If you want different outcomes, you can't just buy a better black box. You have to fix the data the box learns from. AI sits downstream of every choice your company made for decades. Asking it to produce fair results from biased data is like asking a copy machine for an original. The solution is boring but doable:
- Clean your hiring data at the source
- Implement regular reviews to catch bias where it appears
- Audit your ATS to ensure outcomes are fair across protected groups
- Treat hiring algorithms with the same rigor as financial controls
Conclusion: We Can Teach the Machine a Better Way
The Workday ruling is a warning to all vendors and employers. You cannot treat a hiring algorithm as a black box. We never should have treated human-led processes as a mystery either. Somewhere out there is the next Derek Mobley, waiting for a decision made before anyone read his resume. We taught the machine how to make that decision. We can teach it a better one.
Comment