EU AI Act - May 2026: High-risk deadline extended to Dec 2, 2027 (Digital Omnibus); obligations unchanged.

Read the update →

Industry — HR Technology

EU AI Act Annex III explicitly names hiring and HR management AI as high-risk. Here is what that means for your product.

If your product uses AI to rank candidates, assess employee performance, predict attrition, or inform promotion decisions and you have EU users, you are operating a high-risk AI system under the EU AI Act. Belto maps your specific obligations.

Conformity Assessment

High-risk AI systems must undergo conformity assessment before being placed on the EU market. This requires technical documentation, risk management records, and in some cases third-party audit.

EU AI Act — Article 43

Transparency Obligations

Individuals subject to AI-assisted hiring or HR decisions must be informed that AI is being used. The system's purpose, logic, and the individual's right to contest must be documented and accessible.

EU AI Act — Article 13

Human Oversight Requirement

High-risk AI systems must be designed to enable effective human oversight. Decisions cannot be fully delegated to automated systems without a documented human review mechanism.

EU AI Act — Article 14

Common Findings

What Belto identifies in HR AI compliance scans

01

No candidate transparency notice

AI-assisted hiring systems must inform candidates that AI is being used in the assessment process, the logic involved, and how to contest decisions. Most HR AI products deployed in the EU lack this disclosure.

EU AI Act Article 13 · GDPR Article 22

02

Missing human oversight mechanism

High-risk AI systems must be designed to allow effective human oversight. A manager reviewing an AI recommendation without the ability to override, query, or understand the logic does not constitute adequate oversight under Article 14.

EU AI Act Article 14

03

No conformity assessment on file

Before placing a high-risk HR AI system on the EU market, providers must complete a conformity assessment procedure. This requires technical documentation, testing records, and in some cases third-party audit.

EU AI Act Article 43

04

Algorithmic impact assessment absent

Canadian federal regulations require an algorithmic impact assessment before deploying automated decision systems affecting individuals. HR AI systems are explicitly in scope.

Canada ADM Directive · Section 6

Deadlines

Compliance timeline for HR AI operators

February 2025

AI Literacy obligations active

Providers and deployers of all AI systems must ensure sufficient AI literacy for staff involved in operation and oversight. No grace period.

December 2027

Annex III high-risk HR AI obligations

The deadline for full application of high-risk obligations under Annex III — including employment and HR systems — is December 2, 2027, per the Digital Omnibus timeline. Conformity assessments, technical documentation, and human oversight should be completed before that date. Article 4 (AI literacy) and Article 5 (prohibited practices) have been in force since February 2025.

Ongoing

Post-market monitoring

High-risk AI system operators must maintain continuous post-market monitoring. Performance tracking, incident reporting, and annual review obligations are active from the moment of deployment.

Belto maps your HR AI system against every applicable obligation.

From candidate ranking to performance assessment, Belto identifies exactly what your product needs before your EU customers ask.