Purpose
This template provides a cautious, non-conclusive response format for applying LUMINA-30 after an AI incident.
Template
This incident may be examined through the LUMINA-30 boundary question: whether meaningful human refusal authority remained effective before irreversible AI impact.
At this stage, no conclusion is made about fault, liability, or legal responsibility.
The review should first establish:
whether humans could refuse, stop, or override the relevant AI-driven action;
whether that refusal pathway existed before irreversible impact;
whether the pathway was operationally effective rather than merely nominal;
whether sufficient records exist to verify the pathway;
whether AI-generated output became the sole or primary rationale for final action;
whether the AI system, AI agent, platform automation layer, or privileged system had authority to initiate, approve, or execute account recovery, credential changes, tool execution, production actions, rollback suppression, or other irreversible operations;
whether a human actor could stop the AI-mediated action before credentials changed, tools executed, production state changed, recovery options were removed, or rollback became unavailable.
If the refusal pathway cannot be verified, the case should not be treated as demonstrating effective human refusal for LUMINA-30 boundary-review purposes.