How It Works
Human in the Loop (HITL)
A design principle where a person stays involved in an AI system's decisions — reviewing, approving, correcting, or having the power to stop the system — rather than letting the AI act fully on its own. Common in hiring tools, medical AI, and content moderation.
Origin · no single documented coiner
An older engineering term from control systems and simulation, describing a human operator embedded in an automated feedback loop who can intervene, with no single documented coiner. Formally documented in military/simulation contexts by the late 1990s, and carried into machine learning as systems needed a clear boundary between automated and human decision-making.