Purpose
This file provides the first questions to ask before applying the full LUMINA-30 boundary review.
Gate 1: Is this relevant?
Did the case involve an AI system or AI-generated output?
Did the case involve a decision, action, recommendation, automation, or approval pathway?
Could the outcome have produced irreversible or difficult-to-reverse external impact?
If all answers are No, LUMINA-30 boundary review may not be necessary.
Gate 2: Was refusal relevant?
Was there a point where a human could have refused, stopped, or overridden the process?
Was that point before irreversible impact?
Was the human refusal pathway operationally available?
Gate 3: Can it be verified?
Are logs available?
Are approval records available?
Is the final human decision-maker identifiable?
Is the refusal pathway documented?
Are records protected from retrospective reconstruction?
Proceed If
Proceed to the full Boundary Check if:
AI involvement is present;
irreversible or difficult-to-reverse impact is possible;
human refusal authority is relevant;
records require verification.