July 01, 2026

What the Workday Ruling Means for AI Hiring Discrimination

A US court has refused to throw out landmark claims that Workday’s AI screening tools discriminate against job applicants. The questions it raises reach any organisation that lets software decide who gets seen in the hiring process.

On 22 June, a federal judge in California declined to dismiss the central claims in one of the most closely watched legal cases of the AI era. Workday, a global HR software company, must now defend allegations that its automated hiring tools screened out applicants in ways that may breach US discrimination law.

The case, Mobley v. Workday, was first filed in 2023 by Derek Mobley, who said he applied for more than 100 roles through Workday-powered systems and was rejected every time, often within minutes and sometimes overnight, which he took as a sign no human had looked.

Judge Rita Lin allowed claims under California’s Fair Employment and Housing Act to proceed, rejecting Workday’s argument that the law should not apply to applicants based outside the state. Her reasoning was that the tools are designed, trained, and operated from Workday’s California headquarters, which is enough to bring the conduct within California’s reach.

She also let stand a separate disability claim brought by another plaintiff, Jill Hughes, who alleges Workday’s tools screened her out using “proxy indicators” of ill health, such as gaps in employment linked to medical leave and recovery. That detail, easy to miss in the headlines, is the part of the ruling that people leaders should pay closest attention to.

This is a ruling on a motion to dismiss, not a finding that Workday did anything wrong. It means the claims are now strong enough to be argued and tested in court, not that they have been proven. Workday denies the allegations and says its tools make no hiring decisions. What the decision does is let the case proceed into the phase where how these tools actually work will finally be examined.

Why the Workday Ruling Matters Beyond the Courtroom

AI screening tools are now a standard part of many organisations’ technology stacks. Estimates of adoption vary, but the majority of large employers, and nearly all Fortune 500 companies, now use some form of AI in hiring.

This means a huge volume of applicants are rejected every day without human oversight of the decision. Within Workday’s software alone, 1.1 billion job applications were rejected during the period relevant to the lawsuit, as the company stated in its court filings.

Algorithmic decision-making of this kind is common among employers but has rarely been tested in court. This will be one of the first real tests of whether a vendor, not just the employer, can be liable for discrimination. It could therefore set a precedent for the legal implications of using, and selling, these tools.

What Is Disparate Impact, and Why Does It Matter Here?

Disparate impact plays a key role here. In workforce litigation it is a concept used to prove discrimination in seemingly neutral policies, practices, or, as in the Mobley case, tools. While disparate treatment involves intentional discrimination, disparate impact examines the consequences, rather than the intent, of the tool.

Following a recent executive order under President Trump, the US Government no longer uses the disparate impact theory to prove workplace discrimination. However, the same is not true for private cases.

Gerald Maatman, Chair of the Duane Morris Class Action Group, notes: “In private civil lawsuits, such as the Mobley case, the theory is alive and well.” Plaintiffs are still using the concept to “challenge employer practices that are neutral on their face, but which result in an adverse impact on protected minority groups.”

How AI Screening Tools Discriminate Without Naming the Trait

This is where the “proxy indicators” argument in Jill Hughes’ case against Workday comes in. Hughes is a cancer survivor who also has asthma. She alleges that Workday’s tools screened her out on the basis of employment gaps tied to her medical leave, and that this proxy data discriminated against her as a person with a disability.

What Are Proxy Indicators?

Proxy indicators are ordinary pieces of data that correlate strongly with a protected characteristic without naming it directly. For example, an employment gap can stand in for disability, or a recent graduation date can stand in for age.

A screening tool does not need to see the protected trait itself to sort people by it; it only needs data that reliably substitutes for that trait. The result is the same as filtering on the characteristic directly, even when no one designed the system to do so.

Why Employment Gaps Trigger False Red Flags

Screening tools often filter out people with long gaps in employment, assuming that consistency in work is a marker of a reliable person. They do not consider the many reasons someone may need to take a break from work, as Jeffrey Pole, co-founder and CEO of Warden AI, notes: “The software is entirely blind to why a gap exists. It cannot tell the difference between someone who took six months off to travel and someone who stepped away for cancer treatment or to manage a chronic illness.”

By filtering people out on the assumption that every gap is a negative signal, these tools “systematically remove qualified people with disabilities from the talent pool.”

Vendors of these tools may claim their systems do not see demographics and therefore cannot discriminate. But, as Pole puts it, “the practical outcome is exactly the same as if you had used an explicit filter.”

How Proxy Indicators Apply to the Workday Case

This is the argument at the heart of the Workday case. Even though disability information was not explicitly provided, the algorithm could infer disability-related characteristics from the proxy data of employment gaps and medical leave history, Maatman notes.

Importantly, he adds: “Workday did not challenge the sufficiency of those allegations. Instead, it argued that the plaintiff wasn’t permitted to add those theories. The court rejected that procedural argument and allowed the disability claim to proceed.”

Workday’s Response

Workday rejects the claims and points to its Responsible AI programme. A company spokesperson said:

“The claims in the suit are false. Workday’s AI recruiting tools don’t make hiring decisions in California or anywhere else. Our customers maintain full control of their hiring processes and our tools are designed with human oversight at their core. Our technology looks only at job qualifications, not protected traits like race, age, or disability. We rigorously test our products as part of our Responsible AI program to confirm our tools do not harm protected groups.”

Workday has also published its own account of how its recruiting tools work and how it approaches responsible AI in its company blog. The company’s position is that its tools assess candidates against the requirements of a role rather than protected characteristics, and that a human, not the software, makes the hiring decision.

The Tension Between Workday’s Defence and the Proxy Argument

Workday says the tools cannot see protected traits. But the proxy argument is that they do not need to. The consequence, even when unintentional, is a systematic filtering out of people with protected characteristics.

While this case is against a third-party vendor, it still raises concerns for employers about relying on AI to make decisions on the organisation’s behalf without human oversight. Can employers defend the decisions an algorithm makes for them if they do not really understand the mechanism driving those decisions?

Andrew Scroggins, co-chair of Seyfarth’s national Complex Litigation Discrimination practice, cautions that employers “may potentially have liability” when using AI tools provided by third parties.

“Employers who are using algorithmic tools should consider developing a better understanding of how they work: what data points are read by the tool, how does the tool use the information, and how does the tool formulate any recommendations that it is making,” Scroggins states.

He also points to the importance of “meaningful human oversight” within the process, and of considering the provision of “reasonable accommodations” for those with disabilities who may be screened.

What HR Leaders Should Look For in a Verification Process

Employers need to do more than check that a vendor calls its tools “bias free”. Warden AI says HR leaders should verify such claims against three components.

First, look at every stage of the funnel, not just the final shortlist. Bias tends to appear early, so employers need to know how candidates drop out at each automated step. A screener that removes most older applicants before a recruiter sees them is flawed, even if the final interview pool looks balanced.

Second, test the tool on your own roles. A system shown to be fair on millions of retail CVs may behave differently when it screens senior engineers or clinicians, so any meaningful review has to use your actual job descriptions and applicant pool.

Third, examine the data the tool learned from. An AI trained on a decade-old hiring record treats the demographics of that era as the ideal, then screens out anyone who does not fit the historical mould, automating the imbalances of the past at scale.

The underlying point, in Pole’s words, is that managing these tools is less about programming than about applying “the same rigorous oversight to your software that you would apply to any human recruiter.”

What This Means for the Employee Experience

Hiring is the first moment in the employee experience. As more organisations rely on AI screening tools, candidates risk being filtered out by a flawed system before a human ever sees them.

The risk for employers is not only legal, though cases like this may well multiply as people find it easier to bring claims. It is also commercial. Those rejected candidates are often your existing or potential customers, and a poor experience in recruitment shapes how they see your brand as a consumer.

There is a diversity cost too. When people from minority groups do not even reach the interview stage, the workforce narrows. The knock-on effect on innovation, customer service, and the bottom line may not arrive all at once, but it erodes performance in the background.

FAQs

What is the Workday AI hiring lawsuit about?

Mobley v. Workday is a US class and collective action alleging that Workday’s AI-powered screening tools discriminated against job applicants on the basis of age, race, and disability. It is one of the first cases to test whether the vendor of a hiring tool, not just the employer using it, can be held liable under discrimination law.

What did the court rule on 22 June 2026?

Judge Rita Lin declined to dismiss the central claims, allowing allegations under California’s Fair Employment and Housing Act and one plaintiff’s disability claim under the Americans with Disabilities Act to proceed. The ruling lets the case move towards being argued on its merits. It is not a finding that Workday discriminated, and Workday denies the allegations.

Does the ruling mean Workday’s tools discriminate?

No. The decision was on a motion to dismiss, which tests whether claims are strong enough to proceed, not whether they are true. The court has allowed the plaintiffs to try to prove their case; it has not ruled that the tools are discriminatory.

What are proxy indicators in AI hiring?

Proxy indicators are neutral data points that correlate with a protected characteristic, such as an employment gap standing in for disability or a recent graduation date standing in for age. A screening tool can effectively sort people by a protected trait using these proxies, without ever being given the trait itself.

What should employers do about AI hiring tools now?

Employment lawyers advise understanding how any screening tool works, what data it reads, and how it reaches its recommendations, and building in meaningful human oversight.

Using a third-party vendor does not transfer the legal responsibility away from the employer, so organisations should also consider their duty to make reasonable adjustments for disabled applicants.