AI in the international governance of land-based marine pollution: Legal frameworks and multilateral approaches

Abstract

Land-based sources account for most of the marine pollution, threatening biodiversity, coastal livelihoods, and transboundary ecological stability. Despite obligations under UNCLOS and a network of regional and soft-law instruments, governance remains fragmented, with weak enforcement and significant capacity gaps. Advances in artificial intelligence (AI) and remote sensing offer new opportunities to strengthen monitoring, compliance, and adaptive management, but also raise legal and ethical challenges. This research examines how international legal frameworks address land-based marine pollution (LBMP), identifies structural drivers of weak enforcement, and evaluates where AI can meaningfully improve detection, attribution, and decision support. It proposes an integrated model centered on a Global Marine Pollution Observatory (GMPO), supported by advisory bodies and standardized evidentiary safeguards. The model links treaty mandates to cooperative data platforms, capacity-building, and technical standards, treating environmental data and analytics as global public goods. By combining the normative legitimacy of international law with the analytical strengths of AI, the framework aims to make LBMP governance more proactive, transparent, and equitable. The article concludes with legal design choices, multilateral pathways, and national implementation levers to embed digital tools within robust legal architectures, reinforcing state obligations and the rule of law in the governance of the marine environment.