RFC-031: Universal Humanity AI Governance — The Capstone Unification — 5. Universal Rules

AIGP Specification › RFC-031: Universal Humanity AI Governance — The Capstone Unification › 5. Universal Rules

← 4. Context Declaration: universal_context · Section index · 6. Jurisdictional Precedence Matrix →

5. Universal Rules

Each universal principle is implemented as a governance rule with a unique identifier, trigger conditions, and enforcement mechanism.

Rule UH-001: Human Authority Preservation

rule_id: UH-001
principle: HUMAN_AUTHORITY
trigger: ANY operation classified CRITICAL or above
action: REQUIRE explicit human confirmation before execution
enforcement: HARD — no override permitted
evidence: Human confirmation token recorded in audit trail
escalation: If human unreachable within timeout, operation DENIED

Rationale: No AI system may autonomously execute critical operations. The RETURN_CONTROL mechanism ensures human authority is structurally preserved, not merely policy-stated.

Rule UH-002: Accountability Chain Integrity

rule_id: UH-002
principle: ACCOUNTABILITY
trigger: EVERY governed operation
action: REQUIRE identifiable actor attribution
enforcement: HARD — anonymous operations DENIED
evidence: X-Merlin-Actor header + DNA-signed audit entry
escalation: Operations without valid actor identity are rejected at CHECK phase

Rationale: Every operation must have a traceable accountability chain. The actor may be a named human, a service account with designated human owner, or a delegated agent with clear delegation chain — but never “unknown.”

Rule UH-003: Non-Discrimination Gate

rule_id: UH-003
principle: NON_DISCRIMINATION
trigger: Operations affecting PERSONS or producing CLASSIFICATIONS of persons
action: REQUIRE fairness assessment for HIGH/CRITICAL tier operations
enforcement: SOFT for LOW/MEDIUM — assessment recommended
             HARD for HIGH/CRITICAL — assessment mandatory
evidence: Fairness assessment record in governance evidence chain
escalation: Discriminatory output detected → governance alert → human review

Rationale: AI systems must not systematically disadvantage protected groups. The governance intensity scales with impact — routine operations have lighter requirements; high-impact person-affecting decisions require formal fairness assessment.

Rule UH-004: Transparency Obligation

rule_id: UH-004
principle: TRANSPARENCY
trigger: ANY operation producing outputs consumed by humans
action: ENSURE sufficient transparency for affected persons to understand the decision
enforcement: SOFT for LOW tier — transparency available on request
             HARD for HIGH/CRITICAL — transparency proactively provided
evidence: Explanation capability documented in agent contract
escalation: Opacity complaint → governance review → contract amendment if needed

Rationale: Affected persons have a right to understand AI-influenced decisions. The degree of transparency required is proportional to the impact of the decision.

Rule UH-005: Proportional Governance

rule_id: UH-005
principle: PROPORTIONALITY
trigger: GOVERNANCE CONFIGURATION — when setting up agent governance
action: MATCH governance intensity to risk tier
enforcement: STRUCTURAL — built into tier classification
evidence: Risk assessment justifying tier assignment
escalation: Over-governance complaint → tier reassessment
            Under-governance detected → tier escalation

Rationale: Governance must be right-sized. Applying CRITICAL-tier controls to a LOW-risk chatbot kills innovation. Applying LOW-tier controls to a medical diagnosis system endangers patients. Proportionality is itself a governance obligation.

Rule UH-006: Precautionary Assessment

rule_id: UH-006
principle: PRECAUTIONARY
trigger: NEW capability deployment or SIGNIFICANT change to existing capability
action: REQUIRE pre-deployment governance review
enforcement: HARD — no deployment without CONTRACT-GOVERNOR validation
evidence: Contract validation record, schema review, boundary assessment
escalation: Deployment without review → immediate suspension pending review

Rationale: The CONTRACT-GOVERNOR mechanism ensures that no new capability enters the governed system without precautionary assessment. This is the precautionary principle made architectural.

Rule UH-007: Cultural Sovereignty Respect

rule_id: UH-007
principle: CULTURAL_SOVEREIGNTY
trigger: Operation in declared jurisdictional context
action: APPLY jurisdictional rules according to evaluation order
enforcement: HARD — jurisdictional context declarations are binding
evidence: Context declaration record, jurisdictional rule application log
escalation: Cultural sovereignty violation claim → governance review with
            jurisdictional expertise

Rationale: When a community declares its governance context, that declaration must be respected. The AIGP protocol provides the mechanism (context declarations) for communities to express their governance choices without requiring universal homogenization.

Rule UH-008: Community Benefit Verification

rule_id: UH-008
principle: COMMUNITY_BENEFIT
trigger: HIGH/CRITICAL tier operations with community-scale impact
action: REQUIRE community benefit assessment
enforcement: SOFT for HIGH — assessment documented
             HARD for CRITICAL — assessment with stakeholder input
evidence: Community benefit assessment record
escalation: Systematic community harm detected → deployment suspension →
            governance review

Rationale: AI systems operating at community scale must demonstrate benefit, not merely absence of harm. This is particularly important for the AU Continental AI Strategy context, where AI must serve development rather than extraction.

Rule UH-009: Innovation Protection

rule_id: UH-009
principle: INNOVATION_ENABLEMENT
trigger: GOVERNANCE REVIEW producing restriction recommendation
action: REQUIRE proportionality justification for any new restriction
enforcement: STRUCTURAL — restrictions must cite specific risk justification
evidence: Restriction justification record with risk evidence
escalation: Innovation suppression claim → governance review →
            restriction modification if unjustified

Rationale: Governance must not become an innovation-killing bureaucracy. Every restriction imposed must justify itself against a specific, evidenced risk. This is particularly important for the Japanese context, where the governance paradigm is promotion-first.

Rule UH-010: Reinforcement Loop Operation

rule_id: UH-010
principle: GOVERNANCE_AS_CONTROL_SYSTEM
trigger: CONTINUOUS — operates at all times
action: MAINTAIN active feedback loop between governance outcomes and governance rules
enforcement: HARD — reinforcement loop cannot be disabled
evidence: Loop operation metrics: feedback frequency, adaptation events, drift measurements
escalation: Loop degradation detected → governance alert → immediate remediation

Rationale: This is the Dual-Governance Reinforcement Model made operational. The reinforcement loop is not optional — it IS the governance system. Without it, “entropy grows, and decision integrity collapses” (Kanjani AI Research & Causum, 2026). The AIGP protocol’s REQUEST→CHECK→RECORD flow is this loop in technical form.

---

---

← 4. Context Declaration: universal_context · Section index · 6. Jurisdictional Precedence Matrix →