Common Questions

What is Annex III of the EU AI Act and which AI systems does it cover?

Annex III of the EU AI Act (Regulation 2024/1689) lists eight categories of high-risk AI systems that are permitted but subject to strict safeguards. Category 4 covers AI used in employment decisions including recruitment, performance management, and termination.Source: EU AI Act, Official Journal of the EU, August 2024

Do AI layoff tools used by tech companies need to comply with Annex III?

Any AI system that makes or substantially influences employment decisions — including layoff selection — for EU-based employees falls within Annex III's high-risk category. Compliance requires human oversight, technical documentation, and worker access to explanations of adverse decisions.Source: EU AI Act Article 6 and Annex III, point 4

What is the difference between Annex I and Annex III in the EU AI Act?

Annex I lists AI practices that are entirely prohibited (e.g. social scoring, real-time biometric surveillance in public). Annex III lists high-risk AI systems that are legally permitted but must meet strict transparency and human oversight requirements.Source: EU AI Act (Regulation 2024/1689)

Background

Annex III is a schedule to the EU Artificial Intelligence Act (Regulation 2024/1689), which entered into force in August 2024. The Annex lists the categories of AI systems that are classified as 'high-risk' under the Act — meaning they are subject to the regulation's most demanding requirements for transparency, human oversight, technical documentation, and accuracy testing. Annex I of the Act covers prohibited AI practices; Annex III covers high-risk permitted uses that require compliance. The distinction is fundamental: Annex III AI systems can be legally deployed, but only with extensive safeguards.

Among the eight categories in Annex III, point 4 covers 'Employment, workers management and access to self-employment', including AI systems used for recruitment screening, CV filtering, performance monitoring, promotion decisions, task allocation, and termination assessments. Any AI system that makes or substantially influences such decisions and is used in the EU falls within Annex III's scope.

The Annex III employment category has direct relevance to the 2026 AI jobs wave: every major tech company conducting AI-assisted workforce restructuring with EU-based employees must, in principle, ensure that the tools used comply with Annex III requirements — including human review of AI recommendations and worker access to explanations of adverse decisions.

The EU's Digital Omnibus legislative package, agreed provisionally in May 2026, dropped the binding employer AI-literacy duty and softened other compliance requirements, but did not exempt Annex III's employment provisions from their underlying obligations. The Council of the EU gave the Digital Omnibus final adoption on 29 June 2026, confirming that the Annex III recruitment and employment compliance Deadline moves from 2 August 2026 to 2 December 2027, a sixteen-month extension. This creates a regulatory asymmetry across The Atlantic: while the Trump administration's National Policy Framework directs US federal agencies to preempt state-level AI labour laws, EU employers restructuring around AI remain subject to Annex III's human oversight requirements, now with a longer Runway to comply. The compliance gap between US and EU AI employment regulation remains among the most commercially significant divergences of the 2026 regulatory environment.