What Is Human-in-the-Loop?

Human-in-the-Loop — a design pattern in which a human reviews, approves, or can override an AI system's output or action before it takes effect.

Human-in-the-loop (HITL) is the strongest form of human oversight, distinguished from human-on-the-loop (monitoring without approving each decision) and human-in-command (overall control without per-decision involvement). The EU AI Act's Article 14 requires effective human oversight for high-risk AI but does not always mandate HITL specifically. Effective HITL must counter automation bias — a human who rubber-stamps every AI recommendation provides oversight in name only.

Source: EU AI Act, Article 14; ISO/IEC 42001