Artificial intelligence has increasingly become a standard part of how companies recruit and hire. Tools now screen résumés, analyze video interviews using facial and voice recognition, and assign candidates automated scores meant to predict job performance. 

The AI hiring technology market grew to more than $3 billion in 2025, with most employers using some form of automated screening. The efficiency is real, but so are the legal risks. Many business leaders are deploying these tools without fully understanding their exposure.

The Legal Risks Are Not Hypothetical

The same anti-discrimination laws that govern human hiring decisions apply when a machine makes or shapes them. The Equal Employment Opportunity Commission (EEOC) has made clear that an employer can be held liable under Title VII if an algorithmic tool produces a “disparate” or “adverse” impact on protected groups, unless the employer can show the tool is “job related and consistent with business necessity.” 

Critically, an employer may be responsible even when the tool was designed and administered by an outside vendor. The EEOC has issued parallel guidance under the Americans with Disabilities Act (ADA), warning that AI assessments can unlawfully screen out applicants with disabilities. 

State laws are expanding quickly. For example, Illinois enacted the Artificial Intelligence Video Interview Act, which requires employers to notify applicants when AI analyzes their video interviews, explain how the technology works, and obtain consent before use. Effective Jan. 1, 2026, Illinois went further: an amendment to the Illinois Human Rights Act made it a civil rights violation to use AI that has a discriminatory effect on protected classes and barred the use of zip codes “as a proxy” for protected characteristics. 

How Bias Creeps In

The key danger is algorithmic bias. AI tools learn from historical data, so they can absorb and then amplify the very discriminatory patterns employers are trying to eliminate. Emerging academic research has documented racial bias in widely used AI models, including large language models. Because these patterns are buried in code rather than stated in a policy, they are easy to miss and hard to defend after the fact.

Liability and Enforcement Are Accelerating

Enforcement is no longer theoretical. The EEOC sued one company over an AI tool that allegedly rejected older applicants automatically. More significantly, in a recent case, a federal court allowed discrimination claims to proceed against an AI vendor under an “agent” theory of liability, and in 2025 conditionally certified a nationwide collective action for applicants age 40 and over. The case shows how both employers and their software vendors can be drawn into court. 

Practical Steps to Reduce Risk

Counsel and business leaders can take concrete measures:

• Audit your tools for adverse impact before and during use, and document the results.
• Scrutinize vendor contracts, because using a third party does not transfer your liability — seek bias-testing data, audit rights, and indemnification.
• Maintain human oversight so that a qualified person, not the algorithm, makes the final decision.
• Provide required notices and consent where state law requires it. 

The Bottom Line

AI can streamline hiring, but it can also embed discrimination and expose your organization to costly litigation. Before deploying any AI hiring tool, involve legal counsel early — review the technology, the vendor, and your compliance obligations. A modest investment in legal review now is far less costly than defending a class action later.