US States Mandate Bias Audits for AI Hiring Tools Revealing Significant Disparities

In 2025, several US jurisdictions expanded regulatory oversight of AI in employment decisions, building on New York City's pioneering Local Law 144 from 2023. Illinois enacted comprehensive AI bias audit requirements for any algorithmic decision-making tools used in hiring, while Colorado extended its existing AI transparency laws to cover employment screening. New York City expanded its audit requirements to cover internal promotions and performance evaluations beyond initial hiring. The mandated bias audits, conducted by third-party auditing firms, revealed troubling patterns across multiple AI hiring platforms. HireVue, Pymetrics, and several other major vendors showed significant disparate impact against Black and Hispanic candidates, with selection rates 15-30% lower than white candidates for equivalent roles. Gender bias was particularly pronounced in technical positions, where AI screening tools favored male candidates at rates exceeding 2:1 in some cases. Companies using these tools faced immediate compliance challenges. Major employers including Amazon, IBM, and Unilever were forced to suspend AI-powered recruiting in affected jurisdictions while conducting emergency audits. The Illinois Department of Labor issued $1.2 million in fines to companies that failed to complete required audits by the March deadline, while Colorado imposed $800,000 in penalties for continued use of non-compliant tools. The regulatory enforcement revealed fundamental flaws in how AI hiring tools were developed and validated. Many systems had been trained on historical hiring data that reflected decades of workplace discrimination, effectively automating bias at scale. Vendors struggled to implement bias correction techniques while maintaining predictive accuracy, leading to significant performance degradation in some modified algorithms. Legal challenges emerged as affected candidates filed class-action lawsuits alleging civil rights violations. The Equal Employment Opportunity Commission issued guidance supporting state-level bias audit requirements and signaled potential federal enforcement actions. Industry groups pushed back against the regulations, arguing they created an impossible compliance burden and would force companies to abandon beneficial AI tools entirely.