Bots Make the Call, Employers Take the Fall: Liability in A.I. Hiring

Artificial Intelligence (“A.I.”) has quickly become a double-edged sword, streamlining complex decisions and driving productivity while encouraging reliance that can expose employers to liability in ways they may not anticipate. One of the more recent examples of this shift appears in hiring practices. Employers now use automated systems to perform functions traditionally handled by internal hiring teams or outside recruiters, including screening applicants, evaluating candidates, and shaping interview decisions. While this transition offers efficiency in managing large volumes of applicants, it also creates a less visible consequence: employers may rely on these tools without fully understanding how decisions are made, even as they remain responsible for the outcomes.

The scale of modern hiring makes that shift almost inevitable. The volume of job applications has surged dramatically in recent years, with LinkedIn reporting nearly 9,500 applications submitted every minute and individual postings drawing hundreds, sometimes thousands, of applicants within days.[1] As a result, the shift toward automation is not only understandable but often necessary. Today, approximately 52% of hiring managers use A.I.-driven tools to screen candidates, and those systems can process thousands of applications in minutes while reducing hiring time by as much as 30 to 40 percent. These efficiencies are significant. But as adoption expands, so too does the likelihood that employers will rely on these tools at scale without fully understanding how decisions are made.

Liability in A.I. Hiring

This shift reframes a familiar legal question in a new context. The issue is no longer just whether an employer’s decision is efficient. The question is whether that decision can be tied to legitimate, job-related criteria and withstand scrutiny under existing law. A.I. does not replace those obligations. It simply changes how decisions are made and how difficult they may be to evaluate after the fact.

One of the clearest areas of risk is disparate impact.[2] A.I. tools often rely on facially neutral inputs, including education, prior experience, location, or gaps in employment, but those inputs can yield outcomes that disproportionately exclude protected groups. The legal analysis, however, remains unchanged. A process that appears neutral on its face may still violate the law if its effects are discriminatory. The focus, as always, is on outcomes, not intent.

Mobley v. Workday

Recent litigation underscores how quickly these risks are becoming concrete. In Mobley v. Workday, Inc., No. 3:23-cv-00770 (N.D. Cal. 2023), plaintiffs allege that A.I.-driven hiring tools embedded in a widely used recruiting platform systematically screened out applicants across multiple protected classes. The case has already survived motions to dismiss, and the court has permitted disparate impact claims under multiple federal and state equivalent statutes, including Title VII, the Age Discrimination in Employment (“ADEA”), and the Americans with Disabilities Act (“ADA”), to proceed and has been described as a bellwether for how courts may assess A.I.-driven bias in hiring more broadly. Just as notably, the court allowed claims to proceed under the theory that a vendor providing the technology could function as an agent of the employer. That point matters. It reinforces that liability may extend to multiple actors, but it does not relieve employers of responsibility for the decisions themselves.

State Legislatures

At the same time, state legislatures are beginning to respond to these risks. Although Pennsylvania has yet to adopt a comprehensive framework, several jurisdictions have enacted or proposed legislation regulating automated decision-making in employment. States such as Colorado and Illinois have enacted measures specifically regulating the use of A.I. in employment decision-making to address risks of discrimination, while others, including New York and Texas, have adopted broader A.I. governance frameworks that emphasize transparency, safety, and oversight rather than targeting hiring practices directly.

Others, such as California’s proposed “No Robo Bosses Act” (Senate Bill 7), reflect a more aggressive approach. Although vetoed by Governor Gavin Newsom, the bill would have prohibited employers from relying solely on automated decision-making systems (ADSs) to discipline, demote, or fire employees without meaningful human review and corroboration.[3] Legislators have since reintroduced similar measures, signaling continued momentum in this space. On May 1, 2026, the Connecticut House voted to pass the Connecticut Artificial Intelligence Responsibility and Transparency Act (Senate Bill 5), requiring employers to provide notice when A.I. tools materially influence employment decisions and making clear that reliance on such tools does not shield employers from liability.

Key Takeaways

These developments reflect a broader trend toward increased accountability. Employers may be inclined to treat A.I. vendors as responsible for the design and operation of these tools. From a liability perspective, however, that distinction is unlikely to be meaningful. If an employer uses A.I. to screen or evaluate candidates, the resulting employment decision remains attributable to the employer. The emerging regulatory and litigation landscape reinforces that principle. Outsourcing the process does not outsource responsibility.

As this area continues to evolve, employers should take a more deliberate approach when integrating A.I. into hiring processes. At a minimum, employers should understand what inputs their A.I. tools rely on, whether those inputs are closely tied to legitimate, job-related criteria, and how outputs are generated. Regular audits for disparate impact, clear documentation of decision-making processes, and meaningful human oversight are critical. Employers should also evaluate vendor relationships carefully, including contractual protections, transparency obligations, and access to information necessary to assess how the system functions. As courts and legislatures continue to refine the boundaries of A.I. use in employment, those who treat these tools as black boxes risk finding themselves unprepared to justify the outcomes they produce.

If you have questions about implementing A.I. in hiring practices, evaluating potential risks, or ensuring compliance with evolving legal requirements, contact one of  Obermayer’s labor and employment attorneys for guidance.

[1] See also Carolyn Crist, Amid Shifts in Online Recruiting, Job Boards Hold Strong, HR Dive (Sept. 9, 2025), https://www.hrdive.com/news/online-recruiting-job-boards-hold-strong/759595/.

[2] But see Executive Order, Restoring Equality of Opportunity and Meritocracy (Apr. 23, 2025) (directing federal agencies to deprioritize enforcement of disparate impact liability and signaling a shift in federal enforcement priorities, while leaving private litigation and statutory liability unchanged).

[3] See also H.B. 2094, 2025 Reg. Sess. (Va. 2025) (vetoed Mar. 24, 2025) (proposing a comprehensive regulatory framework for “high-risk” A.I. systems); Benjamin W. Perry & Lauren N. Watson, Virginia Governor Vetoes Artificial Intelligence Bill HB 2094: What the Veto Means for Businesses, Ogletree Deakins (Mar. 28, 2025), https://ogletree.com/insights-resources/blog-posts/virginia-governor-vetoes-artificial-intelligence-bill-hb-2094-what-the-veto-means-for-businesses/; Dean Mirshahi, Most Artificial Intelligence Legislation in Virginia Was Tabled, VPM (Feb. 23, 2026), https://www.vpm.org/generalassembly/2026-02-23/virginia-ai-bills-hayes-maldonado-salim-trump-spanberger (noting that the majority of proposed A.I.-related bills in the Virginia 2026 General Assembly session were shelved or continued to 2027 and did not advance).

The information contained in this publication should not be construed as legal advice, is not a substitute for legal counsel, and should not be relied on as such. For legal advice or answers to specific questions, please contact one of our attorneys.