Human-in-the-loop (HITL)

Definition

A design in which a person participates directly at one or more points in an AI-enabled process, often by reviewing or approving an output or action.

Put simply

The AI cannot complete the whole process without a human stepping in.

What boards and leaders should understand

A human approval step can be an effective control, but only if the person has the time, information, competence and authority needed to make a genuine decision.

In high-volume processes, human approval can also become impractical, result in fatigue or turn into rubber stamping.

Human-in-the-loop should therefore be a deliberate control design, not a default reassurance that a person is “in there somewhere”.

Questions to consider

  • Where does human approval genuinely reduce risk, and where might it simply create the appearance of control?