EU AI Act high-risk uses in employment: what candidates should know

The EU AI Act treats AI used to recruit, select, promote, dismiss, allocate work to or monitor people as high-risk. Employers using it must provide human oversight, keep logs, inform workers' representatives and tell people when it is used in decisions about them. Candidates should know those protections, and the HR oversight jobs they create.

What does the Act say about AI in employment?

The EU AI Act sorts AI uses by risk. Its list of high-risk uses includes, in substance, AI systems used:

• to recruit or select people, including placing targeted job adverts, filtering applications and evaluating candidates;
• to make decisions affecting the terms of work relationships, promotion or termination;
• to allocate tasks based on individual behaviour or personal traits;
• to monitor and evaluate the performance and behaviour of workers.

Separately, emotion recognition in the workplace is among the Act's prohibited practices, with narrow exceptions for medical or safety reasons. The prohibitions applied from 2 February 2025.

High-risk status brings duties on both sides. Providers must meet design requirements, including human oversight under Article 14 and record-keeping under Article 12. Deployers, meaning the employers using the system, must follow Article 26. Whether a given agent counts as high-risk is discussed in is an AI agent a high-risk AI system under the EU AI Act, and the category itself in what is a high-risk AI system.

What must employers do, and what does that mean for candidates?

Under Article 26, an employer deploying a high-risk system in employment must, in substance:

1. use it according to the provider's instructions;
2. assign human oversight to people with the competence, training and authority to carry it out;
3. make sure input data it controls is relevant and sufficiently representative;
4. monitor the system and keep its logs;
5. inform workers' representatives and affected workers before putting it into use at the workplace;
6. tell people when such a system makes or assists decisions about them.

GDPR adds to this. Under Article 22, people have the right not to be subject to decisions based solely on automated processing with significant effects, subject to exceptions and safeguards, including the right to human intervention and to contest the decision. A rejected application can be such a decision.

For candidates, the practical points are these. You can expect to be told when a high-risk AI system is used in decisions about you. You can ask what role it played. Where a decision was solely automated, you can ask for a person to look at it. And a person with real authority should be overseeing the system.

Which oversight roles does this create?

The duties above need people. In HR and recruitment, they create roles such as:

• Recruitment reviewer, who reviews candidates an AI system has filtered or scored, with authority to override.
• HR operations analyst, who monitors the system's outputs across groups and keeps logs.
• People analytics or HR systems owner, who answers for one system's scope, data and oversight.
• Employee relations specialist, who handles challenges from candidates and workers.

Titles vary. Read adverts for duties such as screening review, candidate quality checks, fairness monitoring, works council liaison and appeals.

These roles need judgement and a record. Every override should be documented, which is why candidates aiming at oversight work should understand what a decision record is; see what is a decision record and why should job seekers know it.

How can you show employers you could oversee hiring AI?

What might an interviewer ask?

If you are applying for an HR or recruitment operations role, expect:

1. "The screening agent ranks a candidate low because of a gap in their CV. What do you do?" Read the application; gaps have many legitimate reasons. Decide on the evidence and record why.
2. "How would you know if the system was disadvantaging a group?" Compare outcomes across groups and look at which inputs drive the difference.
3. "A rejected candidate asks whether AI was used. What do you tell them?" The truth, in plain words, and how to ask for review.

If you are a candidate being assessed, you can ask the employer, politely, how AI is used in their process and who reviews its results. A reasonable employer will answer.

What does the Act not cover?

It does not ban AI in hiring; it sets duties for high-risk uses. It does not apply to every HR tool; simple scheduling or administrative tools may fall outside the high-risk list. It does not certify individuals. And it does not replace employment, equality and data protection law, which continue to apply. Agents used in HR can also be attacked, for example through manipulated CVs; reviewers should know the risks described in OWASP Top 10 for LLM applications: what every AI agent reviewer should know.

This is an explanation of the rules, not legal advice; confirm how they apply to you with counsel.