The Minimal Decision Record: What a Consequential AI Decision Record Must Contain

This paper presents an applied decision-record schema derived from the Identity-Persistence Program. It asks what a consequential automated decision record must contain so that an independent reviewer can determine what happened without reconstructing missing context. The proposed minimal decision record contains a scope block, an event block, and a completeness audit. The scope block includes at most twelve fields, including a structured operating envelope, and eleven under outcome-equivalence review. The paper distinguishes records from logs: logs report observed behavior, while records declare the governing conditions under which a decision was made. It also distinguishes adjudication-grade records from prediction-grade systems by separating what the system produced from what the declared scope demanded. The paper introduces no new formal theorem; instead, it applies prior formal results to a practical schema and leaves the empirical question—whether such records reduce dispute cost and review time at scale—to pilot validation.