Published on This Week in AI

AI Hiring Tools and Employment Discrimination Law in 2026

Employers are using AI at every stage of the hiring process — resume screening, psychometric assessments, video interview analysis, and automated ranking systems. This adoption has outpaced the legal frameworks regulating it, but 2025 and 2026 have seen significant regulatory and enforcement action in both the United States and the European Union.

How AI Bias in Hiring Creates Legal Exposure

AI hiring tools can produce discriminatory outcomes through several mechanisms. Training data that reflects historical hiring patterns can encode past discrimination: if a model is trained on resumes of people who were hired, it may learn to deprioritize resumes from demographic groups that were underrepresented in past hires, even without any explicit demographic variable. Proxy discrimination is also common: a model that penalizes certain zip codes, educational institutions, or vocabulary patterns may effectively screen out protected groups without directly referencing race, sex, or national origin.

Under US employment discrimination law, what matters is both intent (disparate treatment) and effect (disparate impact). An AI hiring tool that produces statistically significant adverse impact on a protected group violates Title VII even if the employer did not intend to discriminate. The EEOC's 2023 guidance on AI and automated systems confirmed that employers cannot avoid liability by outsourcing the discriminatory step to a vendor's algorithm.

US Regulatory Framework

EEOC Enforcement

The Equal Employment Opportunity Commission has identified automated employment decision tools as an enforcement priority. Its guidance requires employers to: conduct adverse impact analyses on AI tools before deployment; ensure the tool is job-related and consistent with business necessity if adverse impact exists; and maintain human review capability for decisions that significantly affect applicants. The EEOC can investigate AI hiring tools as part of broader discrimination charges and has brought enforcement actions against employers whose AI tools produced documented adverse impact.

State-Level Laws

Illinois' Artificial Intelligence Video Interview Act (effective 2020) requires employers to notify applicants before using AI to analyze video interviews and prohibits relying solely on AI analysis to determine whether an applicant proceeds. New York City Local Law 144 (effective July 2023) requires bias audits of automated employment decision tools before use and mandatory disclosure to candidates. Colorado, Maryland, and several other states have enacted or are considering similar requirements. The patchwork of state laws creates significant compliance complexity for employers operating across multiple jurisdictions.

EU AI Act Treatment of Hiring AI

The EU AI Act classifies AI systems used in recruitment and selection as high-risk AI under Annex III. This means employers and AI vendors operating in the EU must treat hiring tools with the same compliance rigor as medical devices or critical infrastructure. Specifically, high-risk requirements apply to: AI used to sort, filter, or rank job applicants or candidates; AI used to assess candidates during the interview or evaluation process; AI used to make or inform promotion, dismissal, or contract decisions.

Obligations for EU Employers

Under the AI Act, deployers of high-risk AI in hiring must: ensure the AI system is used in accordance with the provider's instructions; implement appropriate human oversight — meaning a human must be able to understand, verify, and override the AI's output; monitor the system for drift and unexpected outputs; maintain logs of AI-assisted decisions for the required retention period; and report serious incidents to the national market surveillance authority. The Act also requires meaningful information disclosure to candidates when AI is used in their evaluation.

Algorithmic Transparency and Candidate Rights

Both EU and US frameworks are moving toward requiring employers to disclose when AI is used in hiring and to give candidates meaningful information about how decisions are made. The EU AI Act's transparency obligations for high-risk AI systems, GDPR rights around automated decision-making (Article 22), and state laws in the US create overlapping disclosure obligations. The practical implication is that employers should document their AI tools, conduct regular bias audits, and be prepared to explain — in human-readable terms — how AI-assisted decisions are made when asked by candidates or regulators.

Frequently Asked Questions

Can a vendor be liable for discriminatory AI hiring tools, or only the employer?

In the US, courts have generally held that employers are the respondent in discrimination claims, not the vendor. But EEOC guidance and the AI Act both push vendors into a compliance role: under the AI Act, the provider of a high-risk AI system carries substantial obligations regardless of who ultimately deploys it. US contract law and vendor agreements are increasingly allocating shared liability for discriminatory AI outcomes.

What is an adverse impact analysis for an AI hiring tool?

An adverse impact analysis compares selection rates for protected groups against the selection rate for the majority group using a statistical measure — commonly the four-fifths (80%) rule: if the selection rate for a protected group is less than 80% of the selection rate for the highest-performing group, adverse impact is presumed. For AI tools, this analysis should be conducted on actual outcome data from the tool as deployed, not just on the training dataset.

Does the EU AI Act apply to non-EU employers hiring EU-based workers?

Yes. The AI Act's obligations apply when the output of an AI system is used in the EU, regardless of where the provider or deployer is established. A US company using an AI tool to screen applications from EU-based candidates is subject to the Act's high-risk requirements for that tool.