EU AI Act HR Compliance Checker (Regulation 2024/1689) — High-Risk AI in Employment

Find out how the EU AI Act (Regulation 2024/1689) classifies the AI you use to hire and manage people. Most recruitment and HR systems land in the high-risk tier, with the main deployer obligations applying from 2 August 2026 and penalties reaching up to €35M or 7% of global turnover. Tick the tools you use to see what's prohibited, what's high-risk, and the duties you'll owe.

✓

Which AI tools does your HR function use?

Screening or filtering job applications (CV/resume parsing, shortlisting)
Annex III(4)(a) Scoring, ranking or automated assessment of candidates (incl. video interviews)
Annex III(4)(a) Placing targeted job advertisements to specific groups
Annex III(4)(a) Decisions on promotion or termination of work relationships
Annex III(4)(b) Allocating tasks based on behaviour or personal traits
Annex III(4)(b) Monitoring or evaluating worker performance and behaviour
Annex III(4)(b) Inferring emotions of staff (affect/sentiment recognition at work)
Article 5(1)(f) Scheduling, payroll or HR chatbots with no decisioning about people
Limited / minimal risk

Your classification

Select the tools you use

No tools selected yet. Tick any that apply to see your obligations and deadline.

Deployer readiness (Article 26)

0 / 7

If you run any high-risk HR AI, these deployer duties apply from 2 August 2026. Tick what you already do.

We assigned competent, trained people with authority to oversee the AI system (human oversight). We informed workers’ representatives and the affected employees before putting the system into use. We use the system strictly in line with the provider’s instructions for use. We monitor how the system operates and keep the system-generated logs for at least 6 months. A human reviews AI-assisted HR decisions and can override them. We completed a fundamental-rights impact assessment (FRIA) where one is required. We have a process to suspend use and notify the provider/authority if the system poses a risk.

Which HR AI counts as high-risk

Annex III, point 4 of the AI Act covers employment, workers management and access to self-employment. Two groups of HR AI are high-risk: systems used to recruit or select people (placing targeted job ads, filtering applications, evaluating candidates), and systems used to make decisions affecting the working relationship — promotion and termination, allocating tasks based on behaviour or personal traits, and monitoring or evaluating performance and behaviour.

Separately, inferring emotions of staff in the workplace is prohibited under Article 5(1)(f) from 2 February 2025, except for genuine medical or safety reasons. A ban is stricter than high-risk: such systems may not be placed on the market or used at all.

Key dates

Date	What applies
1 Aug 2024	AI Act entered into force.
2 Feb 2025	Prohibited practices apply — including emotion recognition at work. AI-literacy duty begins.
2 Aug 2025	Rules for general-purpose AI models, governance and penalties apply.
2 Aug 2026	Full high-risk obligations apply to Annex III systems — including HR & recruitment AI.
2 Aug 2027	High-risk rules for AI in regulated products take effect.

Frequently asked questions

Is recruitment and HR AI really "high-risk" under the EU AI Act?

Yes. Annex III, point 4 lists AI used for recruitment or selection (targeted ads, filtering applications, evaluating candidates) and for decisions on promotion, termination, task allocation, and monitoring of performance or behaviour as high-risk. These carry the Act’s heaviest obligations once they apply on 2 August 2026.

Are any HR AI uses outright banned?

Yes. Under Article 5(1)(f), AI that infers the emotions of a person in the workplace is prohibited from 2 February 2025, except where it is used for medical or safety reasons (for example, detecting fatigue for safety). This ban applies regardless of company size.

We only buy AI tools, we don’t build them. Do the rules still apply?

Yes. As the organisation using the system you are a "deployer". Article 26 puts duties directly on deployers: ensure human oversight, use the tool per the provider’s instructions, monitor it, and keep logs for at least six months. Providers (the developers) have separate, additional duties.

Do we have to tell employees we use the AI system?

Yes. Article 26(7) requires deployers who are employers to inform workers’ representatives and the affected workers before a high-risk AI system is put into use at the workplace, following applicable national and collective-agreement procedures.

What is a fundamental-rights impact assessment (FRIA)?

Article 27 requires certain deployers of high-risk AI (notably public bodies and some private operators) to assess the impact on people’s fundamental rights before deployment — who is affected, the risks of harm, and the human-oversight and mitigation measures in place.