EU Delays High-Risk AI Employment Rules Until December 2027

The European Union has reached a provisional agreement to postpone enforcement of high-risk AI regulations affecting employment decisions until December 2, 2027, extending the original August 2026 deadline by more than a year.

The delay, part of the Digital Omnibus on AI proposal, was agreed upon by the European Parliament, Council of the European Union, and the European Commission on May 7, 2026. The European Parliament approved the provisional agreement on June 16, 2026, and it now awaits formal approval by the Council before becoming law.

What qualifies as high-risk employment AI

Under the EU AI Act's Annex III, high-risk employment systems encompass a broad range of workplace applications. These include AI tools used to place targeted job advertisements, analyze and filter job applications, and evaluate candidates during recruitment. The designation also covers systems that make decisions about employment terms, promotions, terminations, task allocation based on individual behavior or personal traits, and monitoring or evaluating worker performance.

The AI Act employs a four-tier risk framework: unacceptable risk (banned), high risk, limited risk, and minimal risk. Employment-related AI falls squarely into the high-risk category when used for these purposes.

Compliance requirements remain unchanged

While the timeline has shifted, the substance of the regulations remains intact. Employers using high-risk AI systems must notify workers about their deployment, provide human oversight with the ability to intervene when necessary, monitor for potential discriminatory impacts, maintain AI system logs, and ensure compliance with data privacy requirements.

Organizations relying on vendor-supplied AI systems for employment purposes must also follow vendor-provided instructions, which vendors are legally obligated to supply under the Act.

Why it matters

The extended deadline suggests European regulators recognize the complexity of implementing comprehensive AI governance frameworks across diverse employment contexts. For multinational employers and HR technology vendors, this breathing room offers a critical window to audit existing AI deployments, establish compliance protocols, and train personnel—but it doesn't eliminate the obligation. Companies that delay preparation risk scrambling to meet requirements as the 2027 deadline approaches, potentially facing operational disruptions or regulatory penalties.

Timeline and next steps

Once the Council formally approves the amendments, they will enter into force three days after publication in the Official Journal of the European Union. Publication is expected before August 2026.

Despite the extended deadline, legal experts suggest employers may want to continue preparing for the original August 2026 timeline as a precautionary measure. The delay represents a postponement, not a repeal, and the comprehensive nature of the compliance requirements means implementation will require substantial time and resources.

These details were first reported by Ogletree Deakins in their analysis of the provisional agreement.