Beyond the Uniform Guidelines: Why employers should continue monitoring for disparate impact

Reprinted with permission from the July 23, 2026 edition of Daily Report © 2026 ALM Media Properties, LLC. All rights reserved. Further duplication without permission is prohibited, contact 877-257-3382 or reprints@alm.com.

For nearly 50 years, employers and federal agencies have used the Uniform Guidelines on Employee Selection Procedures to evaluate whether selection procedures produce unlawful disparate impact. Although the guidelines may disappear from the federal rule book, disparate impact liability will not disappear with them.

Under the theory of disparate impact discrimination, facially neutral policies that disproportionately affect members of a protected group are generally unlawful, even if the employer did not intend to discriminate. This differs from disparate treatment discrimination, in which an employer intends to discriminate based on a protected characteristic.

The Uniform Guidelines established methods for validating employment practices, monitoring selection outcomes, and documenting compliance under Title VII. For decades, both the U.S. Equal Employment Opportunity Commission and the Office of Federal Contract Compliance Programs have relied on the Guidelines as their principal enforcement scheme in disparate impact claims and for securing significant monetary settlements from employers.

Today, however, many employers are asking whether that framework is obsolete. The EEOC proposed rescinding the guidelines, the U.S. Department of Justice questioned the constitutionality of disparate impact liability, and the Trump administration directed federal agencies to de-prioritize disparate impact enforcement.

It would be easy to look at these developments and conclude that disparate impact is no longer relevant. That conclusion confuses enforcement priorities with governing law. In several respects, particularly the rapid adoption of artificial intelligence tools, the risks have become more significant.

Although federal enforcement priorities have changed, the legal risks associated with disparate impact remain substantial. Employers face a more complex environment than ever before, particularly as AI becomes embedded throughout the employment life cycle, and state and local jurisdictions increasingly regulate automated employment decision-making.

Federal priorities have changed, but the law has not

At the EEOC’s request, the Office of Legal Counsel of the U.S. Department of Justice issued an opinion arguing that disparate impact theory is unconstitutional because it encourages employers to consider race in making employment decisions. The DOJ suggests that the EEOC should modify its standards in evaluating disparate impact claims, such as by making it easier for an employer to show that a challenged practice is necessary for its business. The DOJ’s opinion, however, does not change existing law.

Congress codified disparate impact liability in the Civil Rights Act of 1991, and Supreme Court precedent recognizes it. Unless Congress amends Title VII or the Supreme Court overrules its prior decisions, disparate impact continues to be a recognized theory of liability under federal law.

Rescission of the Uniform Guidelines would eliminate an important source of federal guidance, but it would not eliminate disparate impact claims themselves. Employers would simply lose the long-standing framework that has helped them evaluate their employment practices.

Statistical analysis remains a critical tool

Even if federal agencies pursue fewer disparate impact cases, statistical evidence will continue to play a central role in employment litigation. Plaintiffs’ attorneys routinely rely on statistical analyses when challenging hiring practices, promotion decisions, testing procedures, compensation systems, reduction-in-force selections, and algorithm-assisted employment outcomes. Class actions frequently depend on demonstrating statistically significant disparities for protected groups.

Organizations that periodically evaluate their employment outcomes will be better positioned to identify potential concerns before they result in litigation. Early detection allows employers to investigate whether a practice is job-related and consistent with business necessity and, if not, to explore alternatives. By contrast, waiting for a lawsuit often means identifying potential problems for the first time after a complaint has been filed and the costs of remediation have increased dramatically.

AI makes selection decisions harder to defend

Organizations increasingly rely on technology to make or support employment decisions. Employers are using AI to screen resumes, score recorded interviews, rank applicants, identify candidates for promotion, and forecast workforce needs. These technologies promise greater efficiency, but they also create new legal concerns.

Many automated systems learn from historical employment data. If those historical patterns contain disparities, AI systems may replicate or even amplify them. In addition, many algorithms operate as “black boxes,” making it difficult for employers to explain exactly how recommendations are generated or why certain applicants receive higher scores than others.

Recognizing these vulnerabilities, several states – including California and New Jersey – have enacted or proposed legislation requiring employers to evaluate AI systems for discriminatory outcomes, conduct bias testing, maintain documentation, supply notices, or provide other protections for applicants and employees. Some jurisdictions expressly prohibit the use of AI tools that have a disparate impact unless justified by business necessity. Although the requirements differ across jurisdictions, they share a common concern: ensuring that technology-driven employment decisions do not produce disparate impacts based on protected characteristics.

For employers operating in multiple jurisdictions, compliance has become increasingly difficult. Rather than following a single federal framework, organizations must grapple with a growing patchwork of state and local requirements.

Outsourcing technology does not outsource risk

A common misperception is that purchasing an AI system from a reputable vendor somehow shifts the risk away from the employer. But if a hiring tool screens out applicants in protected groups at disproportionate rates, plaintiffs will generally focus on the employer who used the tool to make the challenged employment decisions, not the vendor or developer.

Thus, employers should understand how their AI systems function, what data they use, and whether independent validation has been performed. Vendor contracts should also address legal compliance, documentation and data retention obligations, cooperation during investigations and audits, and appropriate indemnification provisions. Employers should regularly analyze selection outcomes to determine whether, and to what extent, disparate impact is occurring.

What employers should do now

Reports of the demise of disparate impact are premature. Federal enforcement priorities will continue to evolve, and the legal environment is likely to remain uncertain as courts address new challenges involving AI and the Trump Administration’s position.

Meanwhile, state legislatures are expanding regulation and plaintiffs’ attorneys continue to pursue systemic employment claims. Moreover, employers face increasing expectations from boards, shareholders, and employees regarding the responsible use of AI.

Rescinding the Uniform Guidelines would remove a familiar compliance framework, but it would not erase selection data, the disparities that data may reveal, or the claims that plaintiffs can construct from it. Employers should continue auditing employment outcomes, particularly when automated tools influence who advances and who does not.

To access a recording of our webinar on this topic, click here.