Skip to content

RFC-027: IHL-Aligned Governance — Protected Persons and Humanitarian Context — 1. Problem Statement

AIGP SpecificationRFC-027: IHL-Aligned Governance — Protected Persons and Humanitarian Context › 1. Problem Statement

Section index · 2. IHL Principles Mapped to AI Governance →

RFC-027: IHL-Aligned Governance — Protected Persons and Humanitarian Context

PRIVATE AND PROPRIETARY — NOT A PUBLIC RFC. Owned by Kanjani AI Research & Causum. See NOTICE.md.

Status: DRAFT

Authors: Kanjani AI Research & Causum

Date: 2026-06-17


1. Problem Statement

AI systems increasingly operate in contexts governed by International Humanitarian Law (IHL) — not because they are weapons, but because they process data about, make decisions affecting, or provide services to protected persons in conflict and crisis contexts.

Examples of AI in IHL-governed contexts:

  • Epidemiological modeling during armed conflict (disease spread among civilian populations)
  • Refugee identification and biometric processing
  • Damage assessment and civilian casualty estimation
  • Resource allocation in besieged areas (food, medicine, water)
  • Communication systems serving occupied populations
  • Medical triage assistance in field hospitals
  • Missing persons identification and family reunification
  • Detention monitoring and conditions assessment

In all these cases, the AI system:

  1. Handles data about protected persons (civilians, wounded, prisoners of war, refugees)
  2. May influence decisions affecting protected persons (resource allocation, prioritization, access)
  3. Operates under legal obligations that predate and supersede any governance policy the operator might configure

Current AIGP governance treats all invocations as policy-configurable. But IHL establishes absolute obligations — inviolable floors that no governance configuration may lower. An operator cannot configure a policy that permits the AI to discriminate against protected persons based on nationality, race, or religion (GCIV Art. 27), regardless of what their governance policies say.

AIGP must recognize that some governance constraints are non-derogable — they exist by operation of law, not by operator configuration.



Section index · 2. IHL Principles Mapped to AI Governance →