CIL’s Dr Jon Truby has a new publication with Oxford University Press that tackles a novel question at the intersection of public international law and frontier AI governance: “Global Governance of Transboundary Risks from Agentic AI: Enhancing International Standards for Auditing AI Management Systems to Comply with State Duties” in Oxford Intersections: AI in Society (ed. Philipp Hacker, 2025). DOI: https://doi.org/10.1093/9780198945215.003.0155

The publication explores how public international law duties of due diligence, prevention and cooperation apply to agentic AI systems, which optimise their own actions with limited human oversight and can generate cross-border harm when embedded in critical infrastructure and key sectors. It examines whether existing international standards, particularly ISO/IEC 42001:2023 and the emerging ISO/IEC 42005/42006 series, provide an adequate basis for states to discharge these obligations.

By situating AI management systems within the framework of state responsibility, the chapter connects technical audit standards with international legal duties, and assesses how instruments such as the EU AI Act, the European Commission’s 2025 Code of Practice for general-purpose AI and related ISO standards can support more robust, potentially continuous audit and cross-border supervision of high-risk AI.

The piece argues for a closer integration of public oversight with private certification to manage transboundary risks from agentic AI and to give concrete operational content to states’ existing obligations under public international law.

#PublicInternationalLaw #AIGovernance #AgenticAI