Who's accountable for AI in hiring? What HR buyers need to do now

AI's role in hiring is facing a significant legal test.

The HR technology industry is facing two landmark cases: Mobley v Workday and Kistler et al v Eightfold.

For the first time, liability for AI’s use in hiring is being laid at the door of the developer of technology (the vendor) rather than a deployer (the employer).

The cases directly challenge long-standing positions among vendors that they provide the software but are not involved in how it is used.

These two cases are not uniform – the specific allegations in each case are distinctly different.

Filed in February 2023, the suit against Workday alleges discrimination by its AI technology based on age, disability and race – the US District Court for the Northern District of California has granted conditional approval for the age claims to proceed to a collective action.

Eightfold’s case is a consumer protection action alleging the vendor’s AI scraped the personal data of candidates and employees without the disclosures required by the Fair Credit Reporting Act (FCRA). The suit was filed in January 2026 in a California state court – it was later moved to the US District Court for the Northern District of California, where it is now being heard.

How Workday and Eightfold are responding to the suits

Both Workday and Eightfold reject the allegations and are contesting the suits in the courts.

Eightfold filed a motion to dismiss the claims in April; this motion has yet to be ruled upon by the court.

Workday’s motions to dismiss were partially granted and partially denied. The court dismissed claims of intentional discrimination, but let disparate impact claims proceed.

An Eightfold spokesperson tells UNLEASH: “As a matter of company policy, we do not comment on the specifics of pending litigation; however, we believe the claims are asserted without merit and intend to defend ourselves vigorously.”

“Beyond our formal legal filings, which will speak for themselves, we look forward to addressing these allegations through the appropriate judicial process.”

Both vendors also emphasize their long-standing commitments to responsible AI.

A Workday statement noted: “The claims in this suit are false. Our customers retain full control over their hiring decisions. Our tools are built with human oversight at their core, rigorously tested through our Responsible AI program, and only look at job qualifications—not protected traits like race, age, or disability.”

Kelly Trindel, Chief Responsible AI Officer at Workday, shares with UNLEASH that the vendor has “continued on the same path of developing and improving our responsible AI program, irrespective of the litigation.” In fact, Workday’s program has gained third party certification against the ISO 42001 standard and NIST AI risk management framework.

Eightfold has also prioritized “independent validation” through “third-party bias audits,” including for compliance with the ISO and NYC Local Law 144, the spokesperson shares with UNLEASH.

“Responsible AI is foundational to Eightfold and a core driver of our innovation. Our approach is defined by a rigorous data discipline.”

Not just a Legal wake-up call

Court timelines move slowly – these two suits won't be resolved for months, if not years. But no organization that builds, sells or uses AI in hiring can afford to wait for a verdict before acting.

It may feel natural for HR leaders to park these ongoing lawsuits in the realm of Legal. However, Kyle Lagunas, industry analyst and Founder at Kyle & Co, notes that these cases are highly relevant to HR leaders too.

Both lawsuits “come down to data – what was being captured, and how was it being used?”

“Oftentimes HR looks to our partners in Compliance, Legal and Security to answer questions and make decisions,” but those enterprise-wide data governance approaches may not look specifically at how the data is being used.

Therefore, it’s not sufficient for HR to simply rely on Legal colleagues when it comes to responsible AI.

The same can be said for vendors. Now that vendors may be held liable for their AI tool’s role in hiring decisions, for Jason M. Loring, Partner & Co-Leader of the privacy, data strategy and AI team at Jones Walker LLP, the cases show precisely why “employers cannot assume that responsibility rests with a vendor.”

“The vendor and employer may have separate statutory obligations,” and existing vendor contracts “do not determine whether a statute applies, and they do not erase duties the law assigns to a particular party.”

While “neither case involves employers directly,” Employment Attorney at Bussing Law, Heather Bussing, is clear that “employers are still involved.”

“Vendors will have the bulk of the information about how the programs were designed and configured,” but employers will also “be required to deal with the discovery and human involvement in decision-making.”“Vendors will have the bulk of the information about how the programs were designed and configured,” but employers will also “be required to deal with the discovery and human involvement in decision-making.”

This is coming to fruition in the case Mobley v Workday, where the US court has ordered Workday to disclose which of its customers actually use AI features in hiring.

Ultimately, HR can't afford to wait for a court to define risk — where does that leave the HR buying process?

Don’t abandon, but rethink, the HR buying process

The wrong reaction to these lawsuits is to “stop procuring AI-enabled recruiting technology,” states Loring. Lagunas agrees; there’s no need for HR buyers to distrust vendors.

Switching to building, rather than buying, their HR tools simply “changes the shape of the risk, rather than removing it,” notes Loring. “Neither approach is categorically safer.”

Instead, the focus should be on managing risk around responsible AI use. For Loring, “an organization’s existing diligence processes are the logical place to start.”

Before making any HR buying decisions, leaders need to achieve “decision defensibility,” notes Lagunas. He recommends starting with “internal discovery” on current compliance processes.

HR are not in this alone – Loring says that they should route AI products “through a cross-functional procurement gate.”

“HR, counsel, Privacy, Security, Procurement and appropriate technical personnel” should work together to deeply understand how the tools work before making a buying decision.

Once HR buyers are in a conversation with vendors, “it is not sufficient to ask whether a vendor tested its system.” They cannot just accept “general assurances that a product is accurate, fair, explainable, or supported by responsible AI principles.”

HR teams need to ask vendors better, more specific questions. This includes inquiring “what data is used, how the data is tested for accuracy, whether the software is capable of making employment decisions, and how to use the program effectively without increasing employer risk,” notes Bussing.

This scrutiny extends to contracts. Bussing urges buyers to always read fine print: “The last thing employers want is to get sued and end up paying all of the damages plus the plaintiffs’, the vendors’, and their own attorneys’ fees. Watch for and understand indemnity clauses.”

Lagunas is clear that vendors remain important partners for HR teams – the boundaries just need to be clearer.

Workday’s Trindel has watched that boundary get tested in practice. Trindel has regular conversations with customers, and they will ask directly: What are your recommendations for how to deploy Workday tools? Do you have best practices? “We don’t want to cross the boundary of telling customers what to do, but we also want to be a trustworthy partner.”

For Bussing, HR has a responsibility to audit hiring outcomes: “Keep a close eye on who gets hired and who doesn’t, who gets promoted and who doesn’t, and who is leaving and who isn’t. It’s a good idea anyway.”

Whenever problems are identified – a likely scenario, Bussing admits – investigate the cause, whether it’s the software, recruiters, hiring managers or a combination.

“Document everything and take meaningful steps to mitigate or fix the issue” to avoid any potential liability further down the line.

As HR leaders take another look at their tech stack in light of these suits, Loring’s final advice is not to fall into the trap of viewing the two cases through the same lens. In reality, “an organization can be well-prepared for one while remaining exposed to the other.