Artificial Intelligence and International Law:Human-Centred Approaches to the Geopolitics and Governance of LLMs

Abstract

The rapid development of large language models (LLMs) has generated both enthusiasm and concern in global governance debates. These technologies promise to accelerate legal research, facilitate cross-border cooperation, and access to justice. Yet, their global diffusion has also exposed regulatory gaps, ethical dilemmas, and geopolitical rivalries that transcend national boundaries. The contrasting approaches of the European Union’s AI Act, the United States’ market-driven regulation, and China’s state-centric control illustrate how LLM governance is shaped not only by technical concerns but also by strategic competition. This raises a critical question: how can international law regulate and guide the use of LLMs in a manner that balances state sovereignty, ethical standards, and human rights? This article explores the intersection of artificial intelligence, international law, and geopolitics through a human-centred lens. Using a normative–comparative approach, it examines emerging legal and ethical frameworks such as the EU AI Act, UNESCO’s guidelines on AI ethics, and United Nations initiatives on global AI governance. By contrasting these regimes, the study highlights both convergence and fragmentation, particularly in reconciling innovation with accountability and human dignity. The analysis suggests that while ethical principles are increasingly acknowledged, enforcement remains weak at the international level, and geopolitical rivalries hinder the establishment of coherent norms. The article concludes that international law has the potential to function as a bridge between technological innovation and universal standards of justice. Effective governance of LLMs requires embedding human-centred principles into law and fostering transnational cooperation that mitigates risks, strengthens trust, and promotes global accountability.