The AIHL Project · AI × international humanitarian law

When the fabrication becomes the weapon.

Synthetic media generation now lets belligerents corrupt the records, communications, and identifications that make a captured person legible as a human being owed dignity. AIHL maps where generative AI threatens prisoner-of-war protections — so the harm becomes foreseeable, and therefore preventable.

Geneva law · human dignity GCIII & customary IHL 7-layer threat model 6 analytic steps 3 worked scenarios
The reorientation

From Hague law to Geneva law

The AI-in-conflict debate has matured around the conduct of war. It has been near-silent on how AI renders war more inhumane.

Well-developed
Targeting, proportionality, Article 36 weapons review, escalation
Underdeveloped
Protection of persons hors de combat — POWs, dignity, contact, the record
01

Reorient the discourse

Move from Hague-law concerns about the conduct of hostilities toward Geneva-law concerns: the protection of individual human dignity in captivity.

02

Adapt a threat model

AIHL adapts the MAESTRO agentic-AI framework to assess the harm surface that synthetic media create for POW entitlements under GCIII.

03

Lay the groundwork

Found future technical and policy measures — this platform, red-teaming exercises, and governance mechanisms — to address the vulnerability before harms occur.

Just as much as the laws of war must become a political priority, so too must they become a sociotechnical priority on the part of all AI actors. — On the case for a sociotechnical turn in red-teaming AI systems in armed conflict
The protected interest

The POW information regime

POW protections do not depend on a single record. They depend on a fragile, distributed information regime that must satisfy three conditions to function at all.

V

Validity

Information about POWs — records, communications, identifications — corresponds to reality.

A

Access

Information can be readily reached by the right stakeholders: families, home and detaining states, the ICRC, and the prisoner themselves.

P

Attributability

Information carries clear provenance, allowing it to be traced back to a legitimate actor.

The pipeline · incorrect information cascades through the whole lifecycle
A structural power asymmetry. POWs, their families, and humanitarian personnel systematically lack the technical capacity to verify these conditions — while detaining states have the most capacity, and often the incentive, to fabricate information or disrupt its flow. Defenders must protect both the instance layer (the individual case, where harm is felt) and the systemic layer (the normative power of IHL itself), while attackers need only exploit the instance.
The model

The AIHL framework

MAESTRO's seven-layer architecture maps cleanly onto the pipeline that produces POW-targeted synthetic media — data → model → orchestration → deployment → ecosystem. Six targeted adaptations re-anchor it on humanitarian outcomes and IHL.

The seven layers
What changes · MAESTRO → AIHL
The method · six analytic steps
Interactive walkthrough

Scenario explorer

Select a scenario, then step through the framework as a responsible actor would — decomposing the system, mapping threats across layers, tracing the attack chain, and assigning mitigations.

Distributed defense

Who is the responsible actor?

Decomposing the regime into layers reveals who is best positioned to defend each one. Responsibility is shared and continuous — and it never falls on the victims of harm.

Primary responsibility
None
Low
Moderate
High
Primary
0burden on victims

AIHL places no requirements on the victims of harm, who are least well positioned to defend against synthetic-media violations of their entitlements. Monitors first observe; institutions escalate; platforms interdict; developers harden — each in proportion to capacity, and each able to report safeguard failures upstream.

A qualitative reading of best-positioned responsibility, drawn from the framework. Hover any cell to read the actor × function pairing.
The research

The work behind the project

The AIHL Project rests on a straightforward argument: the protection of people hors de combat deserves the same sociotechnical scrutiny the field already gives to the conduct of hostilities.

Abstract

A sociotechnical turn for the laws of war

Debate over artificial intelligence in armed conflict has matured around Hague law — targeting, proportionality, weapons review, escalation. It has been near-silent on how the same technologies make war more inhumane for those already in custody.

This work reorients that debate toward Geneva law: the dignity, contact, and accurate record-keeping owed to prisoners of war. It shows how generative synthetic media can corrupt the fragile information regime on which POW protection depends, and adapts the MAESTRO agentic-AI threat-modeling framework into AIHL, an instrument that maps the harm surface across seven layers and six analytic steps.

Three worked scenarios — a propaganda deepfake, falsified capture cards, and an audio-cloned call to a prisoner's family — demonstrate the model, and a distributed-responsibility analysis assigns defensive obligations in proportion to capacity, never to the victims. AIHL's central demand is that AI actors treat the laws of war as a sociotechnical priority, not only a political one.

The thesis
On the case for a sociotechnical turn in red-teaming AI systems in armed conflict
Reorienting AI in armed conflict from the conduct of hostilities to the protection of persons hors de combat.
At a glance
7 layers · 6 IHL adaptations
6 analytic steps · 3 worked scenarios
Grounded in GCIII & customary IHL
Reference
AIHL: A threat model for synthetic-media violations of prisoner-of-war protections under international humanitarian law. The AIHL Project, 2026.
Where it goes next
01

Scenario library

An openly navigable, growing library of worked cases, expanding as new synthetic-media threats are analysed.

02

Red-teaming exercises

Operationalizing AIHL as adversarial evaluation against frontier models, with structured probes and scoring.

03

Governance mechanisms

Folding synthetic-media harms into IHL governance, accountability, and cross-actor reporting before harms occur.

Get involved

Pressure-test the protections, before they fail

AIHL is built to be used. It is most valuable in the hands of the actors best positioned to find and close these gaps in their own systems and mandates.

The harm surface here is foreseeable, which means it is preventable. Whether you build the models, safeguard the platforms, hold the humanitarian mandate, or write the rules, AIHL gives you a shared language for naming where synthetic media threatens prisoner-of-war protections — and a method for deciding what to do about it.

AI labs
Model developers & deployers

Harden generative pipelines against the jailbreaks and misuse pathways the scenarios trace.

Platforms
Distribution & provenance owners

Detect and interdict synthetic POW media before it reaches families, states, and the record.

Humanitarian
Institutions & monitors

Anchor evaluation on humanitarian outcomes and escalate failures of the information regime.

Policy
Policymakers & regulators

Bring record-integrity and dignity harms explicitly into IHL governance and accountability.

How engagement works
1

Scope

A briefing maps AIHL onto your systems, mandate, or portfolio and identifies the layers most exposed.

2

Evaluate

Worked scenarios and red-teaming exercises pressure-test those layers against realistic adversaries.

3

Operationalize

Findings become concrete mitigations and monitoring, assigned in proportion to your capacity.

Bring AIHL to your organization.

Request a briefing, propose a red-teaming engagement, or talk through how the framework applies to your work.