AI in recruitment is high-risk under the AI Act. Here it is auditable.

01

Screening is on the list

Annex III point 4 names recruitment and selection, and singles out targeted job ads, filtering applications and evaluating candidates. Point 4(b) reaches promotion, termination, allocating tasks by behaviour or traits, and monitoring performance. Little of a modern HR stack falls outside it.

02

The bias data cannot leave

Article 10(5) lets you process protected characteristics to detect and correct bias, under state-of-the-art safeguards, and then says that data must not be transmitted, transferred or accessed by other parties, and must be deleted once the bias is corrected. Measuring bias needs an environment nobody else reaches.

03

The candidate can ask why

Article 86 gives a person affected by such a decision the right to a clear and meaningful explanation of the part the system played in it. Article 26(7) makes the employer inform workers’ representatives and the affected workers beforehand. Both are hard to answer about a model you cannot inspect.

04

AI literacy came first

Article 4 has applied since February 2025, ahead of the high-risk regime: whoever provides or deploys the system must ensure their staff reach a sufficient level of AI literacy for their role and context of use. That is training rather than infrastructure, and nobody can sell it to you.