Over the past few years, Artificial Intelligence (AI) tools have rapidly spread across several human resource (HR) processes, including recruitment, hiring, performance management, and compensation. While AI can produce information quickly and improve efficiency and consistency, it can also produce false or misleading results. Without human oversight, these false or misleading results, referred to as AI “hallucinations,” can lead to unintended consequences and may create significant litigation risk under existing employment discrimination laws.
Who’s Using It Now? Who’s Not?
According to the Society for Human Resource Management (SHRM), in The State of AI in HR 2026 (the “SHRM Report”), in a sample of 1,908 HR professionals nationwide, 39 percent have already adopted AI into their HR functions; an additional 7 percent reported that they planned to do so in the near future. Another 23 percent said that AI was being used elsewhere in their organization.
AI adoption correlates with company size. According to the SHRM Report, 60 percent of large organizations have implemented AI tools into their HR functions, while only 33 percent of small and 35 percent of midsize companies have done so. Across all organization sizes, the most common use of AI currently is for recruitment.
The survey also reveals a significant gap between organizations actively deploying AI tools and those with no immediate plans to do so. Across the sample, only about one-third, or 31 percent of respondents, work in organizations with no AI initiatives on the horizon.
Legal Risks and Disparate Impact
As more employers begin using AI-enabled recruiting tools to screen resumes, rank and assess applicants, and make hiring decisions, one of the biggest legal risks they may face is disparate impact. Disparate impact occurs when a policy or practice that seems neutral on the surface disadvantages a legally protected group of people.
AI systems can exhibit disparate impact in three ways:
- Human biases can affect how the AI training data is collected and produced — and that bias can be unknowingly fed into the algorithms that humans create.
- Data bias happens when the AI tool is trained on incomplete or unbalanced data. AI tools trained on historical hiring data, for example, can inadvertently replicate and amplify the existing biases present in that data. A well-publicized example of this phenomenon happened at Amazon. The company stopped using its AI-hiring tool when it was found to penalize resumes that included “female-associated” words. The system had been trained on 10 years’ worth of resumes in Amazon’s files, and most of those resumes came from men.
- Algorithm bias happens when systematic errors in machine learning algorithms produce discriminatory results. It happens when the developer — inadvertently or deliberately — overweighs certain factors in algorithmic decision-making. These results often reflect or reinforce existing socioeconomic, racial, and gender biases.
Even though President Trump has decreed the “elimination of disparate-impact liability in all contexts to the maximum degree possible,” and the Equal Employment Opportunity Commission (EEOC) has decided to discharge all disparate impact discrimination claims to align with the president’s executive order, Title VII of the Civil Rights Act remains firmly in place. Title VII expressly recognizes disparate impact claims based on employment practices that disproportionately exclude protected classes. Thus, while the risk of disparate impact exposure probably has been reduced in recent years, it still exists.
Other than disparate impact and the Uniform Guidelines on Employee Selection Procedures to consider, there are several other federal laws that still apply to AI-driven HR tools. These laws include the Equal Pay Act, the Age Discrimination in Employment Act, the Immigration Reform and Control Act, Title I and Title V of the Americans with Disabilities Act, the Pregnant Workers Fairness Act, the Uniformed Services Employment and Reemployment Rights Act, and the Genetic Information Nondiscrimination Act. These employment discrimination laws may come into play if AI tools introduce unlawful discrimination into an employer’s HR processes.
Data Security and Misuse
AI systems usually require collecting personal data — including names, email addresses, and employment histories — and concerns about data privacy and security are growing wherever these systems are used. Without sufficient data protection protocols, employers risk lawsuits and reputational damage.
Another potential problem involves the improper collection and use of job candidate data. AI tools that gather information from candidate social media profiles, for example, could present lawsuit risks if it can be shown that these details have improperly influenced hiring decisions.
Notable Lawsuits
The EEOC settled its first AI discrimination lawsuit in 2023. In Equal Employment Opportunity Commission v. iTutorGroup, Inc., the EEOC alleged that iTutorGroup, an online tutoring company, programmed its recruiting software to automatically reject certain older applicants. According to the complaint, the system screened out female applicants aged 55 and older and male applicants aged 60 and older before recruiters reviewed their applications, resulting in more than 200 qualified applicants allegedly being rejected because of age and in violation of the Age Discrimination in Employment Act (ADEA). iTutorGroup entered into a consent decree with the EEOC and agreed to pay $365,000 to the group of automatically rejected applicants, develop and adopt anti-discrimination policies, and conduct training.
One of the first major class-action lawsuits in the United States alleging discrimination through algorithmic bias is Mobley v. Workday, Inc. The plaintiff, Derek Mobley, an African American man over the age of 40 with a disability, alleged that Workday’s AI-powered applicant screening tools discriminate based on race, age, and disability in violation of federal and state anti-discrimination laws. The class action has gained significant attention because it could set precedent for AI vendor liability in hiring processes.
On January 19, 2024, the court granted Workday’s motion to dismiss the case, with leave for the plaintiff to amend the complaint. On February 21, 2024, the plaintiff filed an amended complaint, with further details to support the claim. The case is ongoing, with hearings scheduled through early 2027.
Final Thoughts on the Implications of AI in the World of HR
There’s no doubt that AI can be a powerful tool for improving efficiency in HR processes — but only if employers remain vigilant about compliance, transparency, and human oversight.
Employers considering AI-assisted hiring tools should, as part of vendor due diligence, work with legal counsel, IT, and HR to thoroughly evaluate what they are purchasing and implementing. The evaluation should include assessment reports and historical selection rates, based on existing anti-discrimination laws and EEOC guidelines.
Carefully negotiate and draft the indemnity, warranty, liability cap carve-outs, and other risk-allocation provisions of the contract with the vendor. These provisions should obligate the vendor to assume liability for any issues arising from the use of the AI tool that are the vendor’s fault. Employment practice liability insurance (EPLI) may not cover certain AI-related claims unless additional riders are in place. Also, joint liability principles mean that both the vendor and the company can be held accountable.
For employers already using AI in HR, legal counsel should work closely with HR and IT to regularly conduct bias audits. If an audit reveals disparate impact, the vendor may be able to modify AI algorithms, add training data for underrepresented groups, or implement other bias-mitigating techniques. When facing scrutiny from regulators or private litigants, the focus will be on how the AI tools were implemented and how well the employer can explain and defend the decision-making process. Employers who evaluate regularly, document their processes, and monitor outcomes will be better positioned to capture the benefits of AI in HR settings, while also reducing litigation risk.
If you have questions or concerns about the legal requirements and potential risks associated with AI bias, or if you’re interested in fostering greater collaboration across your legal, HR, and IT partners to mitigate bias and reduce liability exposure, don’t hesitate to contact Orr & Reno for assistance.
