Could Executives Who Supply AI Targeting Systems Face Rome Statute Liability? The ICRC Blog's 29 October 2026 Case for Article 25(3)(d)(ii) and the Katanga Test

What the ICRC Humanitarian Law & Policy blog published on 29 October 2026

Rosa Villar Orero, a PhD candidate at the University of Aberdeen, argues in the ICRC's Emerging Voices series that senior executives of companies supplying AI targeting systems could, in principle, be criminally responsible under the Rome Statute when their systems contribute to war crimes. Her route runs through Article 25(3)(d)(ii), a residual form of accessory liability for anyone who intentionally contributes to a crime committed by a group acting with a common purpose, with knowledge of the group's intention. The piece is a scholarly argument published on an ICRC platform and should not be read as an ICRC position.

Why the knowledge standard matters for AI vendors

Aiding and abetting under Article 25(3)(c) requires that the supplier act for the purpose of facilitating the crime, which a commercial seller can usually deny. Article 25(3)(d)(ii) asks only for knowledge, and the author applies the five-element test that ICC Trial Chamber II set out in the 2014 Katanga judgment. On her reading, NGO reports, journalism and ICC proceedings can support an inference about what an executive knew of a client's intentions, and that inference, proved against a specific individual, is what could make AI targeting system providers accessories. Public reporting alone does not establish an individual's knowledge, and reported civilian harm does not by itself establish a war crime. The argument extends Rome Statute accomplice liability to software and has not been tested in any court. This post reads the argument element by element and marks where it is strongest and where it is thinnest.