The Decision Proof Gap

Abstract

AI governance frameworks describe how decisions should be made — but at the moment a consequential decision is executed, the evidence required to defend it almost never exists in a verifiable form. This paper names that structural gap, traces it across regulated sectors, surveys why existing solutions (explainability, logs, policy-as-code, model cards) fail to close it, and proposes the Decision Receipt as the missing primitive: a signed, replayable, portable record bound to the policy in force at the moment of decision.