An Analysis of the Foundations of Criminal Responsibility Arising from the Operation of Artificial Intelligence in Light of Recent Developments in International Law

Abstract

^{Artificial intelligence increasingly performs functions traditionally dependent on human judgment, including military targeting, autonomous transportation, predictive policing, surveillance, medical decision-making, and cyber operations. This development has complicated the attribution of criminal responsibility when harmful consequences arise from systems capable of learning, adaptation, and partially autonomous action. This article examines the legal foundations of criminal responsibility arising from the operation of artificial intelligence in light of recent developments in international law. Using a doctrinal and analytical method, it evaluates technical autonomy, legal agency, causation, foreseeability, mens rea, omission, complicity, corporate fault, and command and superior responsibility. The analysis argues that contemporary AI systems lack the consciousness, moral understanding, legal intention, and capacity to experience punishment required for independent criminal responsibility. Accordingly, AI should generally be treated as an instrument, operational intermediary, or risk-generating technology rather than as an autonomous criminal subject. Responsibility should instead be attributed to the natural and legal persons involved throughout the AI lifecycle, including programmers, developers, manufacturers, deployers, operators, corporate managers, military commanders, and public authorities. The appropriate allocation of responsibility depends on each actor’s authority, control, causal contribution, knowledge, ability to foresee harm, and capacity to prevent or terminate unlawful conduct. The article further demonstrates that international criminal law, international humanitarian law, and international human rights law already provide important foundations for addressing AI-related harm, although technological opacity, distributed decision-making, and fragmented organizational knowledge create substantial evidentiary and enforcement difficulties. Particular emphasis is placed on autonomous weapons systems, meaningful human control, due diligence, transparency, traceability, and organizational accountability. The study concludes that increasing technological autonomy must not lead to decreasing human accountability.}