AI Hiring Regulation Is Here: What the EU AI Act and US Laws Mean for Your Recruiting Stack

For years, AI in hiring operated in a regulatory vacuum: vendors made claims, employers bought tools, and candidates had little idea how they were being evaluated. That era is definitively over. Between the EU AI Act, New York City’s Local Law 144, and a growing patchwork of US state laws, automated hiring tools are now among the most regulated applications of AI anywhere - and every employer using them inherits obligations. Here is the map, and what to do about it.

The EU AI Act: Hiring Is Officially High-Risk

The AI Act classifies AI systems used in recruitment - screening, ranking, evaluating candidates - as high-risk, with the corresponding obligations phasing in through August 2026: risk management, data governance, technical documentation, human oversight, and transparency toward the people being evaluated. Two provisions deserve special attention:

• Emotion recognition in the workplace and hiring is banned outright - in force since February 2025. Tools scoring candidates on facial expression or inferred emotional state are not merely dubious science - in the EU they are now illegal.
• Human oversight is mandatory - automated systems may inform hiring decisions, but the architecture must keep meaningful human control over outcomes.

The US: A Patchwork With Teeth

New York City’s Local Law 144 set the template: employers using automated employment decision tools must commission independent bias audits, publish the results, and notify candidates that automation is in use. State laws in Illinois, Maryland, Colorado and elsewhere add consent requirements for AI video analysis and broader algorithmic-discrimination duties. There is no federal statute yet - which in practice means multi-state employers build to the strictest applicable standard.

What This Means Practically

If you are an employer

• Inventory your stack. Anything that scores, ranks, or filters candidates counts - including features buried inside your ATS.
• Demand evidence from vendors: bias audit results, documentation of what the system measures, and proof that scoring is explainable and overridable. A vendor who cannot show the rubric behind a score is a liability you are importing.
• Keep humans decisive, and document it. Auto-rejection by algorithm is exactly the pattern regulators target. Structured AI evaluation that feeds a human decision is the compliant - and better - design.
• Tell candidates. Disclosure is required in more and more jurisdictions, and transparency measurably improves candidate trust anyway.

If you are a candidate

You are gaining enforceable rights: to know automation is used, to see published bias audits in some jurisdictions, and to expect human review of consequential decisions. Use them.

The Contrarian Take: Regulation Is a Gift

Regulation is not the enemy of AI hiring - snake oil is. The rules now arriving ban the pseudo-science, mandate the audits, and reward exactly the tools that were built honestly: structured, transcript-based, explainable, human-supervised.

The compliance burden falls hardest on systems that cannot explain themselves - emotion inference, opaque ranking, black-box rejection. Systems built on structured rubrics, evidence-linked scoring, and human-in-the-loop decisions - the AIHire.io design philosophy from day one - find the new rules describe what they already do. The regulatory era will consolidate the industry around evaluation that can be defended. That is good for employers, good for candidates, and, frankly, overdue.