Human-in-Command (HIC)

An oversight model in which people or an organization retain authority over the overall activity of an AI system: they decide when it may be used, for which purposes, under which limits, and when it must be stopped or replaced. The focus is governance and accountability, not approving every output.

HIC differs from human-in-the-loop because a person need not participate in every decision, and from human-on-the-loop because it is not limited to monitoring operation. It defines the permitted degree of autonomy and retains veto power over use of the system. Without an owner, stopping criteria, and a real override mechanism, the label becomes an organizational chart only.