Conversations about artificial intelligence in the workplace have historically focused on the people the technology eliminates. However, as companies increasingly go "personnel light" that conversation is about to turn toward the people who signed off on the technology. CHROs, CTOs, COOs, or any executives who approved AI hiring, performance, or workforce tools should now assume that when that tool is challenged, they will end up being the scapegoats.
The case that changed the risk calculus
Mobley v. Workday, Inc. (N.D. Cal., No. 3:23-cv-00770) is the canary in the coal mine. The plaintiff alleged that Workday's applicant screening tools discriminated on the basis of age, race, and disability. The court, considering arguments for dismissal, allowed the case to proceed on the theory that a software vendor can be treated as an agent of the employers that use its tools. The court preliminarily certified a nationwide age-discrimination collective in May 2025, and the record now references roughly 1.1 billion rejected applications. A full class certification motion was filed in September 2026, with a hearing set for March 2027.
The main point that matters for executives: The "agent" theory means liability does not stop at the vendor. The employer that chose, configured, and relied on the tool is in the liability event horizon, and invariably a human who made that choice is the easiest person to blame.
In a slightly different context, plaintiffs have also begun arguing that AI-generated candidate scores are "consumer reports" under the Fair Credit Reporting Act. This theory, if accepted, would impose disclosure, authorization, and adverse-action obligations on every employer using them. On the regulation side, states are now legislating potential paths to executive liability vis-à-vis AI. For example, Illinois amended its Human Rights Act effective January 1, 2026 to regulate AI in employment decisions, and California's Civil Rights Council adopted automated decision-system regulations under FEHA. The patchwork of state legislation is growing faster than most companies' compliance programs.
Why the executive, not just the company, is exposed
When an AI-based bias claim surfaces, a company faces a familiar choice: defend the decision or distance itself from it. Distancing requires a decision-maker to distance from. The executive who approved the vendor, who skipped or shortened the bias audit, who accepted the vendor's representations without independent validation, or who authorized deploying the tool becomes the obvious candidate.
Three mechanisms convert a corporate compliance problem into a personal one:
Termination for cause. Most executive agreements define "cause" to include gross negligence, willful misconduct, or material violation of company policy. An AI deployment that produces a discrimination claim could be characterized as any of those three after the fact, particularly if the company adopted an AI policy the executive allegedly ignored. A for-cause termination typically forfeits severance, accelerates nothing, and may claw back equity. For cause termination still tends to be a rare bird; often times even the most obvious for-cause justifications are conveniently ignored by companies that are looking to get rid of an executive. Instead companies tend to favor less friction and the easier negotiation that arises from a negotiated “voluntary separation.” However, if inherent bias in AI hiring practices becomes a huge source of claims, as it has in the Workday case, expect a more mercenary approach by boards looking to somehow limit their exposure at the expense of whoever it is that allegedly brought the AI on board in the first place or failed to properly oversee it.
Oversight liability. Delaware's “Caremark Doctrine” imposes a duty of oversight on directors, and the Court of Chancery's 2023 decision in the McDonald's derivative litigation extended that duty to officers. Courts have treated "mission-critical" regulatory risks as the core of Caremark. An AI tool that touches every hiring decision in a company is a plausible candidate for the mission-critical label, and an officer who had no reporting system for algorithmic bias is a plausible defendant.
Indemnification gaps. Standard indemnification provisions and D&O policies were not written with algorithmic discrimination in mind. Regulatory investigations by state civil rights agencies, FCRA claims, and derivative suits may fall into coverage gaps or exclusions, and the company's obligation to advance defense costs often disappears the moment the company alleges the executive acted in bad faith.
What executives should negotiate now
None of this is unmanageable, but it must be addressed before a problem arrives. In new agreements, amendments, and separation negotiations, I believe executives should address some or all of these four items, depending on context:
Narrow the definition of cause. Carve out decisions made in good-faith reliance on vendor representations, outside counsel recommendations, or internal compliance sign-offs. Require written notice and a cure period before any cause determination tied to a technology deployment.
Expand indemnification and coverage. Confirm that indemnification and advancement expressly cover claims arising from automated decision systems, including regulatory investigations and derivative actions, and that the D&O policy has no AI or algorithmic exclusion.
Document the governance role. Insist on a written description of who owns AI oversight, what the board has been told, and what bias testing the company has performed. If the executive is “person zero” he or she should build the record now: establish or include vendor diligence, audit results, legal review, and board reporting.
Protect severance from after-the-fact re-characterization. Severance that is contingent on a "clean" departure, or severance with a lookback/clawback provision, invites the company to find a reason the departure was not clean. Negotiate for severance that irrevocably vests on termination, or only allows look-back/clawback in very limited situations. As always, cause should be defined as narrowly as possible and determined through an explicitly detailed process.
The bottom line
Companies are deploying AI in employment decisions faster than the law can define the rules, and the first people to absorb the consequences of that gap will not be the vendors. Rather, it will be the executives who approved the tools. The executives who come through this period intact will be the ones who treated AI governance as a personal contractual issue before anyone else did.
J. Andrew Fine represents senior executives in negotiations over employment agreements, separations, and restrictive covenants. FineCounsel has offices in Durham, North Carolina and West Palm Beach, Florida, and handles cases nationwide. Learn more at finecounsel.law.
This article is for general information only and is not legal advice. Reading it does not create an attorney-client relationship. Executives facing the issues described here should consult counsel about their specific circumstances.
Approved an AI tool, or negotiating your next agreement?
FineCounsel negotiates cause definitions, indemnification, and severance protections for executives in North Carolina, Florida, and nationwide. Contact us for a confidential consultation.
Schedule a Confidential ConsultationAttorney advertising. This article is provided for informational and educational purposes only and does not constitute legal advice. No attorney-client relationship is created by reading this article. Laws and plan terms vary. Do not act or refrain from acting on anything in this article without first consulting qualified legal counsel. Prior results do not guarantee a similar outcome.