The expertise layer · Recourse

The right to a human

When a machine makes a decision that changes your life, you should be able to ask for a person. Not a different machine, not a form, a human who can look again and decide. That claim is starting to have the force of law behind it.

The right to a human is the claim that, for a consequential automated decision, a person can demand that a human decide or meaningfully review it. It is a recourse principle, and it is the consumer-facing edge of the whole argument for human judgment infrastructure. As the EU AI Act and related rules come into force, versions of this right are moving from good practice toward obligation for high-risk decisions. But it is worth building even where no law requires it, because a decision with no human door is a decision people are right not to trust.

Where the law backs it

Data-protection law has long recognized limits on purely automated decisions with significant effects, and as newer AI rules come into force, the expectation of meaningful human involvement in high-risk decisions is hardening. I am careful not to overstate the specifics, which are still settling. The direction is clear enough: a reachable human is becoming a requirement, not a courtesy.

Why it is good design, not just compliance

Because the absence of a human door is exactly what turns an ordinary error into a grievance. A genuine right to a human converts a dead end into a relationship and a complaint into a correction. It lowers the trust tax and it catches the wrong decisions the machine was never going to catch itself.