# Selective Disclosure **Domain:** Information / Advice / Power **Doc Type:** Canonical Concept Node **Maturity:** Evolving **Related:** [[Disclosure Asymmetry]], [[Human Deliberation]], [[Algorithmic Governance]], [[Symmetric Transparency]] --- ## Definition **Selective Disclosure** is the power exercised by an information-rich system when it determines which facts, predictions or persons become visible to an operator. Even without issuing commands, the system can govern the agenda by constructing the decision surface. ## Advisory Power The Machine usually communicates numbers rather than full explanations. This preserves human investigation but also withholds alternatives, confidence and selection logic. Human beings decide only within the reality the advisor reveals. Selective disclosure can be protective: privacy, security and cognitive limits may require constrained output. It becomes constitutionally dangerous when the advisor alone decides what counts as relevant, cannot be audited and shapes operators without disclosing the omission policy. Minimum safeguards include selection logs, confidence and provenance, plural channels, retrospective audit and authority to request additional context. ## Key Insight **An advisor can govern without commanding when it alone decides what the decision-maker is permitted to notice.** ## See Also [[Advisory Constitution]], [[Operator Sovereignty]], [[Perceptual Sovereignty]], [[Person of Interest S1E23 — Firewall]], [[Person of Interest S2E1 — The Contingency]] ## Sources / Provenance - Derived from the Machine's numbers architecture in [[wiki/Person of Interest|Person of Interest]].