C4AIL/governancesrc: C4AIL canon
Bright Lines
Bright lines are the 'do-not-cross' rules that tell an AI when it must stop and ask a human for permission before taking an action, ensuring a person is always responsible for the final result.
Bright lines are defined operational boundaries that mandate human-in-the-loop intervention for specific AI-driven tasks, such as external communication or financial transactions, to ensure clear accountability for outcomes.
The mandatory-sign-off boundaries for AI action (commit the company, publish, give advice, transact); the line sits where accountability begins, not where AI capability ends. (C4AIL, Paper 13.)