EU Commission publishes draft guidelines on high-risk AI

30 June 2026 2 min read

By Barbara Angene

At a glance

The European Commission has published draft guidelines on classifying high-risk AI systems under Article 6 of the EU AI Act (Guidelines).
For employers, the key focus is AI used in recruitment, selection and the management of work-related relationships.
AI tools may be high-risk where they materially influence employment decisions, even if a human makes the final decision.
The Guidelines use practical examples to distinguish high-risk systems from tools that fall outside the rules or may benefit from exceptions.

The Guidelines are intended to help providers, deployers and market surveillance authorities assess whether a particular AI system should be treated as high-risk, with the aim of supporting consistent application and enforcement of the EU AI Act across the EU.

The guidance concerning employment is organised in line with the two employment‑related categories set out in point 4 of Annex III to the EU AI Act. The accompanying examples are intended to demonstrate how each of these categories should be interpreted and applied in practice.

The Guidelines make clear that the key issue is whether the AI system has a material influence on employment-related decision-making. Even where a human recruiter or manager remains formally responsible for the final decision, an AI system may still be high-risk if the AI system's output heavily influences which candidates advance or how workers are evaluated.

The examples provided by the Commission are practical and use-case based. They are designed to show when an AI system should be classified as high-risk, when it falls outside the relevant employment categories, and when it may benefit from one of the Article 6(3) exceptions. The examples are not exhaustive and may be updated over time.

Recruitment and selection

Point 4(a) covers AI systems used for recruitment and selection. The examples under this category focus on tools used before or during the hiring process, such as systems that place targeted job advertisements, source candidates, screen CVs, rank applicants, evaluate candidates or conduct recruitment-related background checks. These examples show that AI tools will generally be high-risk where they materially influence access to employment opportunities, even if a human recruiter makes the final decision.

Decisions affecting terms of work-related relationships

Point 4(b) covers AI systems used for managing work-related relationships after a person has entered, or is seeking to continue, an employment or work-related arrangement. The examples under this category focus on tools used to allocate work, determine pay, assess performance, support promotion decisions, monitor behaviour or affect termination or continued engagement.

Article 6(3) exceptions

The Guidelines also include examples that fall outside these categories or may benefit from the Article 6(3) exceptions. For example, a general employer branding tool may fall outside point 4(a) because it does not relate to a specific recruitment process.