Human-in-the-loop describes a process in which human review, feedback, authorization, correction, or intervention is built into an AI-supported workflow.
It is used to improve quality, manage risk, handle edge cases, and keep accountability with people when automation alone is not appropriate.