Artificial Intelligence and International Law: Developing a Global Governance Framework for Autonomous Decision-Making

Abstract

The rapid advancement of Artificial Intelligence AI has fundamentally transformed the nature of decision-making across governmental, commercial, military, and social domains. From predictive policing and autonomous weapons systems to algorithmic financial trading and judicial assistance technologies, AI systems increasingly perform functions traditionally reserved for human actors. While these developments promise efficiency, accuracy, and innovation, they simultaneously challenge foundational principles of international law, including state responsibility, accountability, sovereignty, human rights protection, and the rule of law. Existing international legal frameworks were designed for human decision-makers and state-centric governance structures, creating significant regulatory gaps when autonomous systems independently generate outcomes with transnational consequences. This article critically examines the interaction between AI-driven autonomous decision-making and contemporary international law. It argues that the existing fragmented regulatory landscape is insufficient to address the complex legal, ethical, and geopolitical challenges posed by increasingly autonomous technologies. Current approaches, characterized by soft-law guidelines, national AI strategies, and sector-specific regulations, fail to provide a coherent global governance architecture capable of ensuring accountability, transparency, and legal certainty. The article identifies key shortcomings in current international legal mechanisms, particularly regarding attribution of responsibility, algorithmic transparency, cross-border data governance, and the protection of fundamental human rights. Drawing upon principles derived from international human rights law, international humanitarian law, international trade law, and environmental governance regimes, the study proposes a novel Global Governance Framework for Autonomous Decision-Making ggfadm. The framework advocates a risk-based, multilayered governance model that combines binding international obligations with adaptive regulatory mechanisms. It introduces the concept of “Shared Algorithmic Responsibility,” under which accountability is distributed among states, developers, operators, and corporate entities according to their respective roles in the AI lifecycle. Furthermore, the framework recommends the establishment of an International Artificial Intelligence Agency IAIA responsible for oversight, certification, auditing, and dispute resolution concerning transnational AI systems.