22-10-2026

The EU wants workers to be able to challenge the algorithm that fired them

An algorithm can already decide whether you keep your shift pattern, hit your targets, or get flagged for underperformance. In some workplaces it decides this without a manager ever reading the file. The system scores you, and the score does the talking.

That is the starting point for a new push from the EU: if artificial intelligence is going to rate, reward or fire workers, the people affected need a real way to fight back. Not a feedback form buried in an HR portal, but a formal route to challenge the decision and get a human to look at it again.

This matters because a lot of AI in hiring and management already runs quietly in the background. It filters CVs before a recruiter sees them. It flags which warehouse picker is falling behind. It scores call centre staff on tone of voice. Workers often don't know the system exists until something goes wrong, and then there is no obvious person to appeal to. The EU's approach is to put a name and a process behind that gap.

For job seekers, this is worth knowing even if you never see the system directly. If you're rejected at the first stage of an application and never hear why, there's a decent chance software made that call before a person did. The right to challenge an automated decision isn't just for people already in a job. It applies to hiring too, which is the stage most candidates never get visibility into.

For employers, the practical question is timing. The rules don't all land at once. Under the latest EU AI Act schedule, following the recent Digital Omnibus adjustments, different obligations phase in between December 2026 and August 2028. That's a longer runway than many expected, which gives HR and compliance teams time to actually build the processes, rather than bolt on a disclaimer at the last minute.

The honest risk is that "longer runway" turns into "we'll deal with it later." A right to challenge an AI decision only means something if there's a working process behind it: someone who reviews the case, a reasonable timeline, and a way to reverse a bad call. Companies that treat this as a legal checkbox rather than an actual process will find that out the hard way when the first complaint lands.

We don't know yet exactly how enforcement will work country by country, or how strict regulators will be in the early years. What we do know is the direction: AI can keep making first-pass decisions about work, but it can't be the last word. For anyone job hunting or managing a team right now, that's the detail worth remembering, not the headline about AI taking over hiring.

← More from the blog