Human in the Loop

HITL is one of three oversight levels defined by the EU's Ethics Guidelines for Trustworthy AI, alongside human on the loop (the system acts while a person monitors and can intervene) and human in command (a person decides whether the system runs at all).

The practical rule is to gate by consequence: reads and reversible actions can run on their own with a log, while anything hard to undo or customer-facing waits for a human to approve, edit, or reject. Gate everything and you get rubber-stamping; gate nothing and you get silent damage.