Mobley v. Workday and the joint-employer turn in AI hiring liability

For most of the last decade, the operating assumption in AI hiring has been: the vendor sells the product, the employer uses it, and the liability runs in two separate channels. The vendor faces contract claims, intellectual-property claims, and the occasional consumer protection action. The employer faces Title VII, ADEA, and ADA claims for any discrimination in the resulting hiring decisions.

Mobley v. Workday is the first major federal case to break that separation. The plaintiff alleges that he applied to over a hundred jobs that used Workday’s screening platform, and that the common cause of his rejections — across all of those employers — was Workday’s system, not the individual employers’ discretionary choices. The court refused to dismiss on the theory that Workday could be liable under Title VII and the ADEA as the agent of the employers using its platform.

The analytical move matters more than the headline.

Why "agent" is the load-bearing word

Title VII reaches employers. It does not, on its face, reach the vendors of employment-related software. The agency theory is the bridge: if the vendor exercises sufficient control over the screening decision, the vendor can stand in the shoes of the employer for the purposes of Title VII liability. The vendor becomes — for that specific function — a kind of co-employer.

This is not a new theory. Joint-employer doctrine has been argued in various forms in the staffing-agency context for decades. Mobley is the first major case to extend it to an AI vendor.

What this changes for the deployer

The naive expectation would be that this is a vendor problem. It isn’t.

It is the deployer’s problem because:

• Plaintiffs will not sue the vendor alone. Vendors are deep-pocketed defendants and the joint-employer theory keeps both parties in the case. The deployer faces continued exposure even when the vendor is the dominant defendant.
• Discovery scope expands. What used to be vendor-side discovery is now also deployer-side discovery. The vendor’s testing artifacts, the deployer’s oversight artifacts, and the contract between them all become discoverable.
• Settlement dynamics shift. Vendors will demand contract modifications that push more risk back onto deployers. Indemnification clauses will tighten or disappear.

What the deployer can do

The defensive playbook is documentation, not contract:

• Maintain a clear deployer-side record of where the AI’s output sits in the decision chain, what oversight is applied to it, and what override criteria the human reviewers use. This is the same set of artifacts that the EU AI Act requires. The work is dual-use.
• Document the bias-testing posture you actually rely on. Not the vendor’s testing — yours. Including whatever LL 144 audit you commissioned and any internal testing layered on top.
• Tighten vendor-supplied documentation requirements at contract renewal. The contract should give you the right to obtain the vendor’s bias-testing artifacts in the event of an active matter, not just in the abstract.
• Coordinate with outside counsel earlier. The cost of a Mobley- shaped defense is materially higher if the documentation is being produced after the litigation hold. The cost is far lower if the documentation is already in your hands.

Where this is going

The case has not resolved on the merits. The legal questions are open. But the operational implications for deployers are not contingent on the outcome: even if Mobley settles or is reversed, the next case will be filed, and the documentation that will defend you against the next case is the documentation we recommend producing now.

The Casework engagement was structured around this defensive posture from the start. We do not litigate. If a Mobley-style notice arrives, we coordinate with your outside counsel under the regulator-response support add-on.