Legal risk in employment has long been a challenge as laws are constantly passed to address inequity in a society ever seeking to create a more even playing field for employees, and each change forces employers to adapt to a new landscape. Today, employers are increasingly utilizing AI to assist in day-to-day employment decisions. Tasks ordinarily conducted by trained H.R. professionals, whether as routine as reviewing resumes and employment applications, or as consequential as hiring decisions, conducting performance management processes, and even monitoring employees’ on-the-job activities are now being performed in part, and in some cases entirely, by AI, an emerging technology whose precise benefits and drawbacks are as of yet unknown.
The surface appeal for employers is obvious: AI promises greater efficiency, data-driven insights, emotionless judgement, and cost savings. But as adoption accelerates, so does legal exposure. AI’s powerful yet unnuanced logic may lead to biased workplace decisions that subject employers to liability under a host of federal, state, and local anti-discrimination and privacy laws — even where the employer did not intend to discriminate.
This article surveys the key legal frameworks implicated by employers’ use of AI in the workplace, examines recent litigation that illustrates the real-world consequences of noncompliance, and offers practical recommendations for mitigating risk.
Discrimination Concerns and AI Bias
The most pressing legal risk associated with AI in employment is its potential to produce discriminatory outcomes. For example, some AI hiring software rates job candidates by comparing their micro-facial expressions to those of current employees, and those who differ may be rated lower, disadvantaging them in the hiring process. The core issue is that AI systems are only as unbiased as the data on which they are trained and the sophistication of the algorithms which power them. In combination, these can lead to subtle latent errors that directly contravene multiple established federal anti-discrimination frameworks.
For one, the long-standing Americans with Disabilities Act (“ADA”) protects persons with disabilities from workplace discrimination. Employers may violate the ADA when AI tools unfairly screen-out qualified disabled individuals. The ADA also requires employers to provide reasonable accommodations to individuals with disabilities, including during the hiring process. Users of AI may run afoul of this requirement if the tools preclude or impede the provision of reasonable accommodations – one possible scenario is an AI-powered interview platform not accommodating a candidate’s need for additional time or assistive technology, unintentionally exposing the employer to liability.
Similarly, Title VII of the Civil Rights Act prohibits employment discrimination on the basis of race, color, religion, sex, or national origin. An employer engages in disparate treatment discrimination when it employs automated decision-making tools to intentionally treat members of a protected class differently. For example, in Equal Employment Opportunity Commission v. iTutorGroup, Inc., the EEOC alleged a violation of the Age Discrimination in Employment Act against a company that used an AI hiring program to automatically filter out candidates over a certain age, resulting in a substantial payment to the rejected applicants.
Even where there is no intentional discrimination, AI systems that unfairly exclude protected groups may violate Title VII under a disparate impact theory. Disparate impact discrimination may arise, for example, if a company uses an automated decision-making tool to review resumes and the tool disproportionately excludes female applicants. Unlike disparate treatment, a disparate impact claim does not require proof of discriminatory intent—only that the tool produces an unjustified adverse effect regarding a protected class.
The EEOC has also cautioned that certain AI capabilities may violate employment discrimination laws, and although it has withdrawn some earlier guidance, its reasoning is instructive. Examples cited by the EEOC include video interviewing software that analyzes applicants’ speech patterns and applies low scores to those with speech patterns attributable to a disability, and monitoring software that incorporates facial recognition technology that is less accurate for darker skin tones, leading to potential bias against minority employees. Notably, employers are liable for discriminatory outcomes caused by AI tools regardless of whether the software was supplied by an outside vendor, as the decision to employ a particular tool is enough to trigger employer liability for that tool’s outputs.
One of the greatest challenges employers face is the opacity of AI decision-making. Unlike human decision-makers who can articulate their reasoning, generative AI systems often operate as “black boxes,” making it difficult — if not impossible — to understand or justify how its decisions are reached. This presents a significant obstacle when defending against discrimination claims, because an employer may struggle to present a legitimate, non-discriminatory reason for an adverse employment action if the logic underlying the AI’s recommendation cannot be explained.
Finally, the use of AI to monitor employees raises additional concerns under federal and state privacy laws. If an employer is monitoring or collecting employee information through AI systems, it must consider if applicable laws require disclosure of that monitoring. Failure to provide adequate notice may expose employers to claims under state privacy statutes, while concurrently eroding workplace trust.
The Federal and State Regulatory Landscape
There is currently no comprehensive federal law governing AI use in employment settings. On January 20, 2025, President Trump issued an executive order revoking President Biden’s Executive Order 14110, “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence” which implemented safeguards for AI’s evolution and directed agencies to develop guidance to ensure AI policies were consistent with civil rights.
Shortly thereafter, the Trump Administration issued Executive Order 14179, “Removing Barriers to American Leadership in Artificial Intelligence.” The EO instructed federal agencies to review and roll back existing AI policies and regulations deemed to be barriers to innovation, and to ensure AI systems are “free from ideological bias or engineered social agendas.” This EO signaled a federal preference for minimal regulation in this space.
In the absence of comprehensive federal AI legislation, state and local governments are rapidly filling the regulatory vacuum. Over 1,500 AI-related bills have been introduced across 45 states as of March 2026. This presents an obvious tension between state and the more permissive federal law.
For example, in New York City, employers that deploy automated employment decision tools in hiring or promotion must conduct annual bias audits for race and gender traits, make the results public, and provide transparency notices about the use of the tool to job seekers. Enforcement, however, has proven difficult. Illinois law goes further, making it a civil rights violation for employers to use AI in recruiting, hiring, promotion, renewal of employment, selection for training or apprenticeship, discharge, discipline, tenure, or the terms, privileges, or conditions of employment when such use leads to discrimination based on a protected class — even if the discrimination is unintentional. Employers are required to provide employees the AI product’s name, the employment decisions it influences, the data it collects, and grant them the right to request a reasonable accommodation. Employers in Illinois must also provide notice to candidates and obtain consent before using AI to analyze video interviews.
In Colorado, employers must provide notice when AI is used in “consequential decisions,” a category that includes decisions to offer or deny employment. Employers in Maryland must obtain applicant consent before using facial recognition in job interviews. And in California, employers cannot use AI in any way that causes discrimination in employment, including in hiring, promotion, and screening. Further, California’s Automated Decision-Making Technology Regulations, which will apply to many employers by January 1, 2027, impose detailed requirements before automated decision-making technology can be used for certain employment actions. As of this writing, the Connecticut legislature has passed a bill regulating the use of AI tools in employment-related decision making, and Massachusetts is considering regulations requiring bias audits.
Even employers in states without a local AI regulatory framework may be exposed to legal challenges for the use of AI in their decision-making. Plaintiffs have already started filing putative class actions against AI software companies whose tools assist with employment decisions. No statutes at present preclude individual or class action lawsuits from being brought against employers for challenges based on legal theories listed above – and as discussed, defenses may be hindered when the AI’s underlying decision-making process is unknown.
Ways to Avoid AI Pitfalls
Given the complexity and rapid evolution of the legal landscape, employers should take proactive steps to mitigate the risks associated with AI in employment.
- Employers should implement clear, written policies governing the use of AI in employment decisions. These policies should define the scope of permissible AI use, establish approval processes for new tools, and assign responsibility for compliance.
- Employers must review applicable state and local laws as they relate to AI use in employment practices and obtain all required consent from applicants and employees.
- Employers should train HR professionals and managers on the appropriate use of AI in employment. Training should cover the tools’ technical capabilities and all legal requirements associated with their use.
- Employers should regularly audit and review AI software for bias to help identify and remediate discriminatory patterns before they result in legal exposure.
- Employers should implement meaningful human review of AI-driven employment decisions and ensure that humans retain decision-making authority so that AI is used to inform — rather than dictate — workforce decisions.
Conclusion
While the integration of AI into employment decision-making offers substantial benefits in efficiency and scale, a robotic application of these powerful new tools will introduce significant legal risks. With the shifting regulatory framework on AI, for now, the onus is on employers to proactively monitor how they are using these systems and to make sure they are compliant with changing laws. Discrimination claims, privacy violations, and a rapidly expanding web of state requirements and potential private litigation mean that employers cannot adopt these technologies without careful planning. By implementing robust policies, training their workforces, auditing their tools, and most importantly, maintaining meaningful human oversight, employers can harness the ever-increasing power of AI while minimizing the legal pitfalls that accompany it.
Originally published in the Summer 2026 issue of USLAW Magazine.
This alert is for informational purposes only and does not constitute legal advice. The outcome of the pending litigation remains to be determined and is not guaranteed.
This information is provided for educational purposes only. It should not be construed or relied on as legal advice. It is not intended to create, and receipt of it does not constitute, an attorney-client relationship. If you have specific questions regarding a particular fact situation, we urge you to consult the authors of this publication or other legal counsel.
