Challenges of Attributing Criminal Responsibility for Artificial Intelligence in Hybrid Warfare: An International Humanitarian Law Perspective

Abstract

The increasing integration of artificial intelligence into hybrid warfare has created significant challenges for the attribution of criminal responsibility under international humanitarian law and international criminal law. Autonomous and semi-autonomous systems can participate in surveillance, target identification, operational planning, and lethal decision-making, while their technical complexity may weaken the traditional connection between human intention and harmful outcomes. This study examines the legal difficulties arising from the use of artificial intelligence in hybrid warfare, particularly the accountability gap created when an autonomous system produces conduct amounting to a serious violation of international humanitarian law without an immediately identifiable human perpetrator possessing the required mental element. The study adopts a descriptive-analytical and comparative legal approach and focuses on the relationship between operational autonomy, mens rea, command responsibility, individual participation, and state responsibility. The analysis demonstrates that artificial intelligence cannot presently be regarded as an independent bearer of criminal responsibility because it lacks human-like intention, knowledge, free will, moral understanding, and the capacity to experience punishment. Accordingly, technological autonomy should not be equated with criminal or moral agency. The study argues that responsibility should instead be reconstructed through a hierarchical human-centered model. At the operational level, military commanders remain responsible for lawful deployment, supervision, and meaningful human control. At the technical level, programmers, developers, and manufacturers may incur responsibility where their intentional or legally culpable contribution satisfies the relevant legal requirements. At the sovereign level, states remain responsible for the deployment and consequences of military artificial intelligence attributable to them. The study further emphasizes the importance of meaningful human control, algorithmic auditability, traceability, weapons review, preservation of operational records, and mechanisms for compensation. It concludes that artificial intelligence should remain legally classified as an instrument of warfare rather than an autonomous criminal subject and that international law should prevent technological complexity from becoming a mechanism for avoiding human accountability.