NTRVSTA / Compliance / European Union

Compliance · European Union

EU AI Act for hiring: compliance checklist

The EU AI Act lists AI used to recruit or select people, including tools that filter applications and evaluate candidates, as high-risk (Annex III, point 4). Two rules already apply: emotion recognition in the workplace has been banned since February 2, 2025, and so has the AI literacy duty. The high-risk duties for employers using these tools were due August 2, 2026, but the 2026 amending regulation (EU) 2026/1744 moved them to December 2, 2027.

Last reviewed November 20, 2026 against the statute and regulator text linked below.

Not legal advice. This page summarizes public law for hiring teams. It is not legal advice and does not create an attorney-client relationship. Laws and rules change; confirm how they apply to you with employment counsel.

EU AI Act at a glance

LawRegulation (EU) 2024/1689 (AI Act), as amended by Regulation (EU) 2026/1744 of 8 July 2026
Recruitment AIHigh-risk under Annex III, point 4(a)
Already appliesBan on emotion recognition at work (Art. 5(1)(f)) and AI literacy (Art. 4), since February 2, 2025
High-risk dutiesApply to Annex III systems from December 2, 2027 (was August 2, 2026)
Who it coversEmployers using the tools ("deployers") and vendors ("providers") in or serving the EU
PenaltiesProhibited practices up to €35M or 7% of worldwide turnover; deployer breaches up to €15M or 3%

Who it applies to

  • Employers that use AI in recruitment are "deployers." Vendors that build and sell the tools are "providers." The heaviest documentation duties fall on providers; employers carry the deployer duties.
  • Annex III, point 4(a) covers AI intended for recruitment or selection, in particular to place targeted job ads, analyze and filter applications, and evaluate candidates. Point 4(b) covers promotion, termination, task allocation and performance monitoring.
  • That reaches AI resume screening, AI interviews that score candidates, and candidate ranking.
  • The workplace emotion-recognition ban applies to any AI that infers emotions of people at work, except for medical or safety reasons.

What employers must do

  1. Stop any emotion recognition now (Art. 5(1)(f))Inferring candidates' or workers' emotions with AI is prohibited, outside medical or safety uses. Check interview tools for features that read emotion from face or voice.
  2. Support AI literacy now (Art. 4)Take measures to support the AI literacy of staff who operate or use AI systems, considering their training and the context of use.
  3. Human oversight (Art. 26(2), from Dec 2, 2027)Assign oversight to people with the competence, training, authority and support to do it.
  4. Data, monitoring and logs (Art. 26(4)-(6))Ensure input data you control is relevant and sufficiently representative, monitor the system per the provider's instructions, and keep the logs it generates for at least six months.
  5. Inform workers and candidates (Art. 26(7), 26(11), 86)Before using a high-risk system at work, inform workers' representatives and affected workers. Inform people that they are subject to a high-risk system that makes or assists decisions about them, and be ready to give clear explanations of the AI's role in individual decisions.

Key dates and deadlines

  • August 1, 2024The AI Act enters into force.
  • February 2, 2025Prohibited practices (including workplace emotion recognition) and AI literacy apply.
  • July 24, 2026Amending Regulation (EU) 2026/1744 published in the Official Journal; consolidated text dated July 27, 2026.
  • August 2, 2026Most remaining provisions apply, including Article 50 transparency.
  • December 2, 2027High-risk duties for Annex III systems, including recruitment AI, apply.
  • August 2, 2028High-risk duties for AI in regulated products (Annex I) apply.

European Union AI hiring compliance checklist

Need the candidate notice itself? The free AI hiring notice generator drafts European Union notice text for your company and role, plus an audit-log checklist.

How NTRVSTA and Carolina help

NTRVSTA does not claim EU AI Act conformity. What it gives an EU hiring team today is the material oversight depends on: scorecards with written reasons, highlights and a person who makes the decision.

Structured scorecards with written reasons

Every AI interview comes back as a structured scorecard with a transcript and highlights, and every resume is scored against your scorecard with the reasoning written down per candidate. That gives you something to explain, not just a number.

A person makes the call

Scores are one input. You choose whether each Carolina action is suggest-only, ask-first or automatic, and offers and comp changes always wait for a person.

Evidence you can keep

Shareable candidate reports include scores, highlights, notes and video snippets of key moments, so a reviewer can see what the score was based on.

A real, published bias audit

NTRVSTA's AI interviews had an independent NYC Local Law 144 bias audit by Asenion, finalized October 20, 2026. The summary is on our compliance page.

No software makes you compliant on its own. Your notices, your decisions and your records are still yours.

Questions

Is AI recruiting software high-risk under the EU AI Act?

Yes. Annex III, point 4(a) lists AI systems intended for the recruitment or selection of people, in particular to place targeted job ads, analyze and filter applications, and evaluate candidates, as high-risk.

When do EU AI Act obligations for hiring tools apply?

The high-risk duties for Annex III systems, which include recruitment AI, apply from December 2, 2027 after Regulation (EU) 2026/1744 amended Article 113. The ban on workplace emotion recognition and the AI literacy duty have applied since February 2, 2025.

Can I use emotion detection in video interviews in the EU?

No. Article 5(1)(f) prohibits AI systems that infer emotions of people in the workplace, except where intended for medical or safety reasons. Candidates in a hiring process for your workplace should be treated as covered.

What does an employer have to do under the EU AI Act?

As a deployer of high-risk AI: use it according to the provider's instructions, assign competent human oversight, ensure input data you control is relevant and representative, monitor it, keep logs for at least six months, inform workers' representatives and affected workers before use, and inform people subject to its decisions.

What are the fines under the EU AI Act?

Up to €35 million or 7% of total worldwide annual turnover, whichever is higher, for prohibited practices such as workplace emotion recognition, and up to €15 million or 3% for breaches of deployer obligations. For SMEs the lower amount applies.

Sources

For a longer walkthrough, read our EU AI Act hiring guide.

‍