Employers that rely on AI hiring discrimination defenses may face a risk that extends far beyond the current pause in federal enforcement.
Legal shifts but private claims persist
In April 2025, an executive order redirected federal agencies away from pursuing disparate‑impact claims. The order created a perception that AI‑driven hiring tools are now low‑risk for large firms. That view overlooks two important points. First, the order does not affect private lawsuits under Title VII, the Age Discrimination in Employment Act, or the Americans with Disabilities Act. Second, future administrations can reverse the policy, leaving today’s liabilities intact for a new government to enforce.
Cases already in the courts illustrate the growing exposure. The most visible example is Mobley v. Workday, Inc., filed in the Northern District of California. The plaintiff, an African‑American man over forty, applied for more than a hundred positions through Workday’s AI screening system and was rejected each time, often within hours. The court allowed the claims to proceed on an agent‑liability theory, rejecting the notion that software decisions are insulated from antidiscrimination law. A collective action covering applicants over forty was conditionally certified in May 2025, and the court later rejected Workday’s argument that the ADEA does not apply to job applicants.
Other filings reinforce the pattern. In Baker v. CVS Health Corp., an applicant alleged that an AI video‑interview platform functioned like a lie detector, a claim that survived a motion to dismiss and settled. The ACLU’s 2025 complaint against HireVue and Intuit accused an AI tool of discriminating against a deaf Indigenous applicant. The EEOC’s first AI hiring discrimination suit, against iTutorGroup, alleged age‑based rejections and settled as well.
Related: How Personal Injury Attorneys Maximize Compensation for Clients
Across these cases, courts have shown a willingness to let claims move beyond the pleading stage, and settlements have become common. Quiet settlements, however, do not generate public precedent; instead, they signal to plaintiffs’ counsel that such lawsuits have monetary value.
From a practical standpoint, this means that companies cannot rely on the belief that AI decisions are objectively neutral. They must build a record that demonstrates proactive steps to assess and mitigate bias.
For workers, this shift could translate into more robust protections when AI filters out candidates, especially older applicants who may already feel sidelined by technology.
Liability remains despite the temporary enforcement lull. State agencies in California, Colorado, Illinois, and Texas are already pursuing AI‑related discrimination claims, filling the gap left by federal de‑prioritization. Moreover, any future administration could revive aggressive enforcement, targeting the same actions taken during the current term.
Related: Court Rules in Favor of OppFi
Steps employers can take now
Legal counsel should advise clients to audit their AI tools. Companies need to identify which systems influence hiring decisions and verify whether those tools have undergone independent disparate‑impact testing. Tools that make final accept/reject decisions carry the greatest risk.
Vendor contracts must be updated. Indemnification clauses drafted before the agent‑liability theory emerged likely do not address the new risk. Representations about bias testing, audit rights, and data handling should become standard negotiating points. Vendors unwilling to provide transparency on testing methods should raise a red flag.
Documentation is essential. Firms should keep records of why each AI tool was selected, what due‑diligence was performed, and any vendor assurances about bias testing. If internal concerns arise and are not acted upon, that documentation will surface during discovery, potentially harming the defense.
Monitor developments regularly.
