From August 2026 the EU AI Act fully applies to AI systems used in recruiting, selecting and assessing people. The act classifies those applications as high risk. That sounds technical, but the core is simple: when AI helps decide about people, a human must demonstrably be able to intervene.
What the act requires
For HR processes, high risk means three things:
1. Transparency. Candidates and employees must know when AI plays a role in a decision that affects them.
2. Bias control. Organisations must be able to demonstrate that their systems do not structurally disadvantage groups. That requires validated instruments and periodic checks.
3. Human oversight. There must be a human who understands the system's output, can weigh it and can overrule it. Not a formality after the fact, but a meaningful role in the decision.
The pitfall: oversight as a checkbox
The biggest pitfall we see: organising human oversight as an administrative step. Someone clicking approve without genuinely being able to judge the outcome is not oversight. The act asks for a professional who knows the instrument, understands the context and can explain the outcome to the person concerned.
Why this is not a turning point for us
At AVOP this has always been the standard. In our assessments technology strengthens the work: validated tests, data and structured observation. But the psychologist remains in charge. Every report is written, weighed and explained by a human who has actually seen and spoken to the candidate. Not because the law requires it, but because a decision about people requires human judgement.
For organisations that have come to rely on fully automated screening and algorithmic selection in recent years, August 2026 will be a turning point. For organisations that kept human judgement central, it is a confirmation.
What you can do now
Map where AI already plays a role in your HR processes, including the quiet use that never appeared on an agenda. Check whether every AI-supported decision includes a meaningful human role. And have the conversation in your team about where the line sits: which part of decisions about people stays human work?
Want to know more about how we look at AI? Read about AI & organisational development, or see how human oversight works in our assessments.