AI and Employment Law in the US: Navigating Uncertainty and Regulatory Change

As AI becomes increasingly common in HR departments, including through automated tools used to identify and screen job candidates, employers must remain aware of their legal obligations. In the United States, there is no comprehensive federal law addressing the use of AI in the workplace. Instead, several US states have passed legislation regulating employers’ use of AI in employment decisions. At the same time, the Trump administration has issued executive orders arguably in conflict with these state efforts. As a result, employers must navigate an increasingly complex legal landscape to make informed decisions based on their jurisdictional reach and risk profile.

Colorado provides one of the most comprehensive examples of state-level regulation in this area. In 2024, the state passed legislation governing the use of AI in employment-related decision-making, with the law scheduled to take effect in late June 2026. As currently drafted, the law would apply to businesses of all sizes doing business in Colorado and would require employers using AI tools for consequential employment decisions to conduct impact assessments designed to identify discriminatory trends. The law would also require employers to provide notice to affected individuals, such as job applicants, and offer opportunities to appeal adverse decisions. The legislation is clearly aimed at promoting diversity.

Since its passage, however, the law has faced significant opposition from AI developers and the broader business community. In March 2026, the Colorado governor proposed a draft bill that would replace the existing one. The following month, Elon Musk’s company, xAI, filed a lawsuit challenging the law’s constitutionality, and the federal government joined the lawsuit at the end of April. The parties, including Colorado, xAI, and the federal government, have agreed that the state will not enforce the law while the court considers preliminary motions.

Regardless of the outcome of the litigation, the dispute highlights a broader trend toward piecemeal state regulation of AI in the workplace. In the absence of comprehensive federal legislation, other states may adopt similar measures modelled on Colorado’s approach. To the extent these laws focus on preventing AI systems from reinforcing historical discrimination against disadvantaged groups, they may conflict with the Trump administration’s focus.

Early in his administration, President Trump issued several executive orders suggesting that the federal government views certain common DEI practices as forms of unlawful discrimination. See our prior blog: Best Practices for Maintaining DEI Programs in the US. In July 2025, President Trump linked those concerns directly to AI by issuing an executive order directing federal agencies to purchase only AI large language models that are “neutral, nonpartisan tools that do not manipulate responses in favor of ideological dogmas such as DEI.” In March 2026, he issued another executive order requiring federal agencies to include specific provisions in government contracts and subcontracts. Among other things, contractors would be required to certify that they do not engage in racially discriminatory DEI activities, as defined by the administration. The order also warned that violations could give rise to liability under the False Claims Act, a federal statute historically used to address fraudulent claims involving government funds rather than DEI programs.

The administration has demonstrated a willingness to enforce this position aggressively. In April 2026, IBM agreed to pay more than $17 million to resolve allegations that it failed to comply with representations made in its federal contracts regarding its DEI practices. This marked the federal government’s first DEI-related settlement under the False Claims Act.

These developments leave employers in a difficult position. On one hand, the federal government is directing contractors to eliminate DEI programs that might, allegedly, impart benefits on certain groups of employees and applicants (for example, through affinity groups or diverse candidate slates). On the other hand, some states are encouraging diversity including by requiring employers to ensure that AI systems do not perpetuate historical discrimination or bias in hiring and employment decisions.

In many cases, employers may be able to navigate this tension by ensuring that DEI initiatives avoid conferring tangible preferences, quotas, or segregated benefits while also taking steps to ensure that AI systems do not produce discriminatory outcomes, such as disproportionately screening out older applicants. At the same time, each company’s risk profile will differ. Businesses with federal contracts or a significant public presence may choose to prioritise compliance with the Trump administration’s interpretation of what amounts to unlawful DEI practices. By contrast, international companies with limited US operations and no federal contracts may find it easier to align with state-level requirements such as Colorado’s, particularly where those rules resemble legislation adopted in other jurisdictions.

Until Congress adopts a unified federal framework governing AI in employment, companies operating in the United States will likely continue navigating a patchwork of evolving and, at times, conflicting legal obligations. Employers should therefore monitor both federal and state developments closely to make informed compliance decisions that support long-term business growth.