Is AI Hiring Software High-Risk Under the EU AI Act? A Classification Guide
Whether an HR tool is high-risk determines most of your obligations under the EU AI Act. The answer depends on what the system is intended to do and how it influences decisions, not on the vendor's marketing.
Short answer
AI hiring software is high-risk under the EU AI Act when it is intended to be used for recruitment or selection, in particular to place targeted job adverts, analyse and filter applications, or evaluate candidates, as listed in Annex III point 4(a). AI used for decisions on promotion or termination, task allocation based on behaviour or traits, or monitoring and evaluating workers is also high-risk under point 4(b). A system may fall outside high-risk status if it only performs a narrow procedural or preparatory task and does not materially influence decisions, but AI that profiles people is always treated as high-risk.
Key takeaways
- Classification is based on intended purpose and actual use.
- Annex III point 4 covers both recruitment and management of workers.
- The narrow-task exception is limited and must be documented.
- Profiling of individuals keeps a system high-risk regardless of the exception.
Annex III point 4: employment
| Point | Covers |
|---|---|
| 4(a) | AI intended for recruitment or selection, in particular to place targeted job adverts, analyse and filter job applications, and evaluate candidates |
| 4(b) | AI intended to make decisions affecting terms of work relationships, promotion and termination; to allocate tasks based on individual behaviour or personal traits; and to monitor and evaluate the performance and behaviour of people in work relationships |
The narrow-task exception
Under Article 6(3), an Annex III system is not considered high-risk if it does not pose a significant risk of harm, including by not materially influencing the outcome of decision-making, for example where it only performs a narrow procedural task, improves the result of a completed human activity, detects decision-making patterns without replacing human assessment, or performs a preparatory task. However, an Annex III system is always considered high-risk where it performs profiling of natural persons. Providers relying on the exception must document their assessment.
A classification method
- Describe the intended purpose: what the provider says the system is for, and how you actually use it.
- Check for prohibited practices: such as workplace emotion recognition.
- Map to Annex III point 4: does it support recruitment, selection or decisions about workers?
- Check for profiling: does it evaluate personal aspects of individuals, such as performance or behaviour? If so, it is high-risk.
- Assess the exception: is it genuinely narrow, procedural or preparatory with no material influence?
- Check transparency duties: does it interact with people or generate content?
- Document the result and review when use changes.
Worked examples
| System | Assessment |
|---|---|
| Tool ranking applicants by predicted fit | High-risk: filters and evaluates candidates; involves profiling |
| Tool converting CVs into a standard format without scoring | Possibly within the narrow procedural exception; assess and document |
| Productivity analytics scoring individual employees | High-risk: monitors and evaluates behaviour; profiling |
| Shift scheduling based on stated availability only | Likely not high-risk; check whether it uses behaviour or traits |
| AI that flags potential bias in managers' ratings for HR review | Possibly within the pattern-detection exception; assess and document |
Examples are illustrative only; classification depends on the specific system and use.
This is general information, not legal advice. The AI Act is being implemented through guidance, standards and national enforcement that continue to develop. Take qualified advice on your specific systems and jurisdictions.
See EU AI Act impact on recruitment and employer obligations.
Related guides
- The EU AI Act and HR: What Employers Need to Know in 2026
The AI Act for HR, updated for the 2026 Omnibus: what is high-risk, what is banned, and the timeline.
- EU AI Act Impact on Recruitment: What Changes for Hiring Teams
Which recruitment tools are high-risk, what changes for hiring teams and candidates, and how to prepare.
- AI Tools for Employee Performance Tracking: Outcome Tools vs Surveillance
Categories of AI performance tools, the line between tracking outcomes and surveillance, and selection criteria.
- AI Resume Screening: How It Works, Risks and Best Practice
How AI reads, matches and ranks CVs, where it fails, and how to use it fairly.
Frequently asked questions
Is all HR software high-risk under the EU AI Act?
No. Only AI systems intended for purposes listed in Annex III point 4, such as recruitment, selection, decisions on promotion or termination, task allocation based on behaviour or traits, and monitoring or evaluating workers, are high-risk.
What is the Article 6(3) exception?
An Annex III system is not high-risk if it does not materially influence decision-making, for example performing a narrow procedural or preparatory task. Systems that profile individuals are always high-risk.
Is employee monitoring software high-risk?
AI intended to monitor and evaluate the performance and behaviour of workers is listed in Annex III point 4(b) and is high-risk. Emotion recognition in the workplace is prohibited.
Who decides whether a system is high-risk?
Providers classify their systems based on intended purpose, but deployers should assess how they actually use a system, since use outside the intended purpose can change the analysis.