A job applicant is rejected by every employer he applies to through AI-screened applicant tracking systems, and he sues. That is the case at the center of this episode of HR Voices, the first in its new scenario format. Host Rebecca Taylor asked Kristen Duckett, Chief Human Resources Officer at WorldStrides, what stood out as most risky. Kristen started with a distinction: “AI tools are not inherently discriminatory.”
Her next point carried the rest of the conversation. A neutral tool does not move the responsibility off the company using it. Kristen, who began her HR career in executive recruiting, sees the real exposure in the governance most companies skip when they rush to adopt AI, not in the algorithm itself.
Why the Lawsuit Is a Governance Question
The case is Mobley v. Workday. According to Fisher Phillips’ summary of the litigation, Derek Mobley, a Black man over the age of 40, applied to more than 100 jobs at companies using Workday’s AI-based hiring tools and was rejected each time. Workday argued it was not the employer making the decisions. In May 2025, the court granted preliminary certification of a collective action under the federal age discrimination statute. Kristen noted that the case was still ongoing when they spoke.
The defendant is the vendor. Kristen’s questions went to the employers anyway: whether they had a steering committee, who decided which tools to use, and whether anyone had trained the recruiting team. She blamed much of the risk on fear of missing out: leaders hear about AI everywhere and implement it before the guardrails exist.
Her fix was to slow down in order to speed up later. In her words, the work is “slowing down to speed up.” That means controlling who can use the tools, understanding how a vendor’s screening algorithm works, and asking whether the vendor runs bias testing. It also means looking back at six or twelve months of results for adverse impact by group rather than “setting and forgetting” the system. A neutral-looking screen can still create a disparate impact claim if it filters out one group at a higher rate.
What Automation Bias Does to a Recruiter’s Judgment
Kristen’s sharpest warning was about people, not code. These tools, she argued, are “meant to inform decisions” and “not meant to be the decision maker.” The trouble is that humans tend to defer to them.
She pointed to University of Washington research on the effect. In the UW hiring study, 528 participants picked candidates for sixteen jobs. With no AI help, they chose white and non-white applicants at equal rates. With a severely biased AI, their choices followed its recommendations about 90 percent of the time. Awareness of the bias did not cancel it.
That finding turns training into a control. Kristen wants recruiters taught to exercise their own judgment and managers checking how rankings are used. Some tools offer explainability features that show why candidates were ranked as they were. A recruiter who sees an odd reason can dig deeper, and that is often how a buried bias comes to light.
What Actually Works
Build a Steering Committee Before You Buy
WorldStrides runs an AI steering committee: Kristen, the general counsel and assistant general counsel, the CIO, and the chief product officer. She called it a “four-legged stool.” Each leg sees risks the others miss. HR should know which tools are in use, who owns each one, and how each vendor tests for bias. That governance comes first; scale comes after.
Train Recruiters to Question the Ranking
Recruiters should not carry the blame for a tool leadership chose. Kristen would not want them to feel the problem is theirs. Shared accountability still grows as teams learn the tools, so a recruiter who spots something wrong should raise a hand. That takes the psychological safety to speak up, and leadership has to build it.
Design a Partial Pause, Not a Full Stop
Lawsuits can run for years, and hiring cannot stop. Kristen’s answer to that pressure was targeted. Stop using the tool to screen out resumes; keep using it to schedule candidates. Scheduling sits among the lower-risk AI use cases for People teams she considers innocuous. Her phrase was to “keep the train running,” because a program pulled entirely is hard to restart. The deeper move is change management for how work gets done: rethink the process first, then use AI to amplify it instead of dropping it on top of old ways of working.
Where Employee Relations Fits
The conversation kept returning to one discipline: investigate before you defend. Rebecca named a reflex common in HR teams that assume the company is right and the claimant is wrong. Kristen agreed that the first job is finding out whether something went wrong. Running a structured workplace investigation answers that question before anyone writes talking points.
Kristen also widened the lens. A bias in hiring raises the question of where else the same data drives decisions, such as career pathing and promotion. One complaint can be the first sign of a systemic pattern across the employee lifecycle. Tracking concerns in a single structured case record makes that pattern visible early, not after a lawsuit surfaces it.
Her caution on messaging follows from the same logic. A company that spins a narrative and is later found culpable loses its credibility entirely. Employees, in her experience, will give a second chance to leaders who acknowledge fault and explain what they will do differently.
Frequently Asked Questions About AI Hiring Governance
It is a discrimination lawsuit over AI screening. Derek Mobley says he was rejected from more than 100 jobs at companies using Workday’s tools. In May 2025, a federal court preliminarily certified an age discrimination collective action in the case.
In Kristen’s view, the head of HR. If the CHRO brings the tool in, budgets for it, and signs off on the vendor, the outcome belongs to the CHRO, whether the cause was a glitch or a poorly governed rollout.
It is the tendency to trust an algorithm’s ranking over your own judgment. University of Washington researchers found people largely followed a severely biased AI’s picks, so recruiters need training to push back.
Kristen recommends HR, legal, IT, and product. At WorldStrides, the committee includes the CHRO, the general counsel and assistant general counsel, the CIO, and the chief product officer, so each function can spot risks the others miss.
Usually not entirely. Kristen suggests pausing the high-risk feature, such as resume screening, while keeping low-risk uses like interview scheduling running. Shutting everything down damages momentum and trust that are hard to rebuild.
The Bottom Line for HR Leaders
Kristen’s framing is simple: the vendor builds the tool, but the company owns how it is used. AI hiring risk is a governance gap, not a technology flaw. The CHRO closes it with a cross-functional committee, trained recruiters, and regular audits of outcomes. When a claim lands, she would investigate first, pause only what is risky, and tell employees plainly what is known.
Her closing advice came from home. “The cover up is worse than the crime,” Kristen said. Leaders who come forward early, and say they are committed to fixing the problem, keep the trust that a defensive response spends.
See how AllVoices helps HR leaders document, investigate, and resolve concerns before they become public problems.
