The EU AI Act's hiring rules moved to December 2027. Here's what already applies to your hiring tools

Juan Molina ยท

The high-risk rules for AI in recruitment were delayed, but some obligations are already in force. A plain-language guide for companies hiring in Europe.

If your company uses AI anywhere in hiring in the EU (screening CVs, ranking applicants, running AI interviews, targeting job ads), you may have heard two conflicting things this year: that the EU AI Act's hiring rules arrived in August 2026, and that they were postponed.

Both have some truth in them. Here's what changed, what didn't, and what it means in practice. (This is a plain-language summary of public legal analysis, not legal advice. Check with your counsel before acting.)

What was delayed

The EU's Digital Omnibus on AI, adopted this summer, moved the obligations for stand-alone high-risk AI systems to December 2, 2027 (Freshfields, July 10, 2026). That category includes AI used in "employment and worker management", which covers most AI recruitment tools. Law firm Jones Walker confirms the December 2, 2027 date for employment obligations, noting that it depends on the amendments' publication in the EU's Official Journal (Jones Walker, July 16, 2026).

So the heaviest requirements for high-risk hiring tools are about 14 months away.

What already applies

Some parts of the Act are already in force, and they matter for hiring:

1. Emotion recognition at work is banned. Since February 2, 2025, the AI Act has prohibited "the use of AI systems to infer emotions of a natural person in the areas of workplace and education institutions," except for medical or safety reasons (AI Act, Article 5). If any interview or assessment tool claims to read candidates' emotions, that's a problem today, not in 2027.

2. Transparency duties from August 2, 2026. Most of the Act's transparency obligations still apply from August 2, 2026 (Jones Walker). In general terms, people should be told when they're interacting with an AI system (Freshfields), which is relevant if candidates talk to a hiring chatbot or AI interviewer.

3. AI literacy. Article 4 asks providers and deployers of AI systems to take measures supporting a sufficient level of AI literacy among the people who operate those systems on their behalf (AI Act, Article 4). For hiring, that means the recruiters and managers using AI tools should understand what those tools do and where their limits are.

What to do now

The delay is time to prepare, not a reason to wait:

  • List every AI tool in your hiring process, including features inside your ATS or assessment platforms that you may not think of as "AI".
  • Remove anything that infers emotions from candidates' faces, voices or writing.
  • Tell candidates when AI is involved. Say it plainly in job ads and invitations, and before any AI interview.
  • Ask your vendors for documentation on how their tools work, how they're tested for bias, and how they'll meet the 2027 high-risk requirements.
  • Train the people who use the tools, so a human can explain, question and override what the AI recommends.

The short version

The EU AI Act's high-risk hiring rules were pushed to December 2027, but the ban on workplace emotion recognition, the transparency duties and the AI literacy duty already apply. Companies that use the extra time to map their tools, be open with candidates and train their teams will be ready when the deadline arrives, and will be fairer to candidates in the meantime.

If you're hiring in Europe and want to talk through how to keep your hiring process both effective and fair, you can book a 15-minute call.

Not ready for a call yet? You can post your open role for free.