The Workday Lawsuit Changed AI Hiring. Here's How to Screen Resumes Without Becoming a Defendant.

A federal judge just put AI resume screening on trial.

In Mobley v. Workday, the U.S. District Court for the Northern District of California ruled that discrimination claims against Workday's AI-powered screening tools can move forward. The numbers are staggering: roughly 1.1 billion applications were rejected through Workday's tools during the relevant period, and the conditionally certified collective could reach hundreds of millions of applicants.

If your team uses AI anywhere in your hiring process — or is thinking about it — this case is now required reading. But most of the coverage is asking the wrong question.

The wrong question: "Is AI biased?"

The instinct is to debate whether algorithms discriminate. The court's concern was simpler and far more practical:

Here's the detail that should reshape how every employer vets hiring technology. Workday's AI never had to "see" a candidate's age, race, or disability status. In most applicant tracking systems, EEO data is walled off from the review process entirely. The legal exposure came from proxy data — signals that correlate with protected characteristics without naming them.

Graduation year. Career length. Employment gaps. Job titles. Compensation expectations.

The court specifically declined to dismiss the claim that the software flagged applicants using proxy indicators of disability — including gaps in employment — in potential violation of the ADA.

The two-year gap problem

Consider the scenario the court zeroed in on. A candidate applies with a two-year gap in their work history. The system ranks them lower, or filters them out before a human ever looks.

Is that disability discrimination?

Maybe. That gap could be cancer treatment. A mental health leave. Caring for a disabled family member. Or it could be a layoff, a sabbatical, or a failed startup. The algorithm doesn't know the difference — it just knows the gap correlates with something it was configured to treat as a negative signal.

The same logic applies to age. When a 25-year industry veteran applies to a role scoped for three years of experience and lands at the bottom of the ranking, is that level fit — or age discrimination against a worker over 40?

Here's the key distinction: when a human recruiter passes on a gapped resume, you can ask them why. When a black-box algorithm does it, the reasoning is buried inside a model nobody can inspect. If you can't see how the system reached its judgment, you can't evaluate whether it was reasonable, flawed, or discriminatory.

That's the whole ballgame in a lawsuit.

"But it's the vendor's problem, right?" Wrong.

The EEOC has been explicit: employers are liable for bias in AI tools used on their behalf — including third-party tools marketed as bias-free. And recent research out of Stanford showed a tool can pass an aggregate bias audit while still systematically screening out protected groups for specific roles.

A vendor's passing audit is not your defense strategy. Your defense strategy is being able to explain every screening decision you make.

The eight questions every hiring team must be able to answer

Whether you're evaluating a new tool or defending your current stack, these are the questions that matter:

1. 1What data does the tool actually see — and what is excluded?
2. 2Does it infer protected characteristics through proxies?
3. 3How are job requirements weighted, and who decided?
4. 4Can it distinguish exact experience from transferable experience?
5. 5Is it rejecting candidates, ranking them, or organizing them?
6. 6Can a recruiter override the recommendation?
7. 7Is there an audit trail for every screening decision?
8. 8Has it been tested for adverse impact at the role level — not just in aggregate?

If your vendor can't answer these, that's your answer.

How HireFab answers all eight — on screen

We built HireFab for small hiring teams that want AI's speed without the mystery box. The design choices map directly to the questions above.

Faster is only better if it's explainable

AI can be genuinely transformative in recruiting, especially when 300 applications land for one role and you need to surface qualified candidates fast. But speed without transparency is how you end up scaling a weak process across thousands of applicants — and into a courtroom.

The Workday ruling makes the standard clear: no hiring decision should be impossible to explain.

That's not a burden. For teams with the right tools, it's a competitive advantage.

HireFab provides decision-support software. Nothing in this post is legal advice, and no software can guarantee compliance with employment law. Consult counsel about your hiring process.