The military deployment of Lethal Autonomous Weapon Systems (LAWS) and the growing reliance on artificial intelligence algorithms in lethal decision-making have confronted the conventional frameworks of International Humanitarian Law (IHL) and International Criminal Law (ICL) with an “accountability gap” and severe challenges in establishing the mental element (mens rea). Employing a descriptive-analytical methodology grounded in international instruments, judicial precedents, and legal doctrine, this study re-evaluates the legal regimes governing liability for international crimes resulting from the operation of such systems. The findings demonstrate that claims of computational autonomy and the “black-box” phenomenon do not preclude legal liability, nor can the deployment of these weapons serve as an accountability shield. On the one hand, rejecting any criminal agency or legal personhood for autonomous machines—and characterizing them strictly as instrumentalities of crime—individual criminal liability attaches to human agents through indirect perpetration (Article 25(3)(a) of the Rome Statute), the doctrine of command/superior responsibility (Article 28 of the Rome Statute), and the technical complicity/aiding and abetting of software developers and weapon suppliers (Article 25(3)© of the Rome Statute, supported by the jurisprudence of the Hague District Court in the Frans van Anraat case). On the other hand, pursuant to the 2001 ILC Articles on Responsibility of States for Internationally Wrongful Acts (ARSIWA), the acts of LAWS are attributable directly to the deploying State through the exercise of governmental authority, operational mission tasking, and breaches of positive obligations—notably pre-deployment legal reviews under Article 36 of Additional Protocol I—entailing State liability for ...