Zeggagh, Maha

This article examines responsibility under international law for military algorithms and autonomous weapon systems. It starts from the growing transfer of parts of combat decision-making to systems that can detect, classify, prioritize, and sometimes select and engage targets. The central question is whether the responsibility gap results from a weakness in international law itself, or from the difficulty of attributing conduct and proving knowledge, foreseeability, and control when an algorithm mediates the use of force. The article adopts a doctrinal and critical legal approach, focusing on international humanitarian law, international criminal responsibility, and state responsibility. It develops the analysis through six sections: the transformation of combat decision-making, the limits of international humanitarian law, the responsibility gap, individual criminal responsibility, state responsibility, and meaningful human control. The article concludes that international law does not face a complete legal vacuum. Its rules remain relevant, but their application becomes more difficult when system behavior is hard to foresee and combat decisions are hard to trace. The article therefore argues for stronger weapons reviews, documented operational constraints, functional transparency, and restrictions on systems whose behavior cannot be reasonably foreseen.