Knowledge MaxxingHuman-on-the-loop
Human-on-the-loop describes a system that acts automatically but is monitored by a person who can intervene or halt it if something goes wrong, rather than approving each action in advance.
Prerequisite: Human-in-the-loop
What matters
This design trades the delay of prior approval for faster automated action, which only stays safe if monitoring is genuinely active and intervention is actually possible in time.
Example
An automated content filter that removes flagged posts immediately, while a moderator monitors and can reverse mistakes afterward, is human-on-the-loop.
Common misconception
Human-on-the-loop is not automatically safer than a fully automated system; if monitoring is too slow or infrequent, harmful actions can still occur before anyone intervenes.
Recall question
What trade-off does human-on-the-loop make compared to human-in-the-loop?