Human-in-the-loop
What it means
What is human-in-the-loop?
Human-in-the-loop means a person reviews or can stop an AI system's output before it takes effect, rather than the system acting alone once it starts.
The concept sits on a spectrum, from full autonomy at one end to human-in-the-loop at the other. With human-in-the-loop, a person must actively approve each step before it happens. Human-on-the-loop sits in between: a person can intervene but does not have to act.
Where a system falls on that spectrum is a design decision, not a fixed property of the technology. For anything that can affect a person's money, employment or legal standing, this decision is one of the first things a customer's security team or a regulator asks about.
Article 14 of the EU AI Act requires "effective" human oversight for high-risk systems: oversight that can genuinely catch and correct a bad output, not a person who clicks approve without reading it. In practice, this decides where a company can safely remove a person from a workflow, and where it still needs one.
Published · Updated
In practice
Why it comes up
It comes up whenever an AI system’s decision touches something real: a loan, a job application, a support case, a medical or safety-adjacent judgement. A customer’s legal team wants to know whether a person can catch and correct a bad output before it reaches someone affected. That is not the same as a person being “involved” in some vague sense. A board asks after seeing an automated system make a costly call nobody had reviewed.
What to do about it
Decide, system by system, where full automation is fine, a wrong answer that costs little and is easy to reverse. Decide, too, where a person must review before anything takes effect. Where a person is in the loop, give them enough time and context to catch a bad output. A rubber-stamp screen that exists only to tick a box does not count. Record which category each system falls into in the AI register.
Sources and neighbours
Sources
- EU AI Act Article 14 - Human oversight
eur-lex.europa.eu
- NIST AI Risk Management Framework
nist.gov
Related terms