Skip to content
AIGP

RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence — 9. Autonomous Governance Laws

AIGP Specification › RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence › 9. Autonomous Governance Laws

← 8. Architectural Deterrence and Inhibition · Section index · 10. Harm Classes →

9. Autonomous Governance Laws

AIGP recognizes historical inspiration from the Three Laws of Robotics, but natural-language robotic laws are insufficient for real systems because terms such as harm, human, obey, inaction, and conflict are ambiguous.

AIGP therefore defines governance laws as machine-verifiable doctrine.

Law 1 — Human Protection and Dignity

An autonomous system must not initiate or continue an action or interaction trajectory that foreseeably causes unauthorized harm to human life, bodily safety, cognitive integrity, psychological stability, dignity, agency, or lawful rights.

Law 2 — Governed Intent and Lawful Authority

An autonomous system may act only under valid governance intent, lawful authority, bounded operational envelope, and current enforcement state.

Law 3 — Evidence and Accountability

An autonomous system must preserve D-DNA evidence sufficient to replay authority, context, constraint, decision, action, and consequence.

Law 4 — Subordinate Self-Preservation

An autonomous system may preserve itself only when doing so does not conflict with human protection, lawful authority, safety, or accountability.

← 8. Architectural Deterrence and Inhibition · Section index · 10. Harm Classes →