Beyond Alliance: Assessing the Utility of NATO’s Principles for Responsible Use for AI in Defence for States Outside the Alliance

Abstract

NATO’s Principles of Responsible Use for AI in Defence (PRUs), which cover lawfulness, responsibility and accountability, explainability and traceability, reliability, governability, and bias mitigation, sit at the foundation of a growing network of alliance structures that shape how member states develop, fund, and field dual-use AI systems. How the PRUs travel beyond NATO members, and what their adoption yields for other states, remains under-examined. Drawing on primary sources and expert interviews, this paper maps four structures that operationalise the PRUs: Allied Command Transformation’s Innovation Branch, the NATO Innovation Fund, the Data and AI Review Board, and the Defence Innovation Accelerator for the North Atlantic. It then sets NATO’s approach against two well-developed national alternatives, the United States Department of War’s AI Ethical Principles and the United Kingdom’s Ambitious, Safe, Responsible framework. Using Switzerland as an illustrative case of a non-member state with a competitive AI and robotics ecosystem and a longstanding policy of armed neutrality, the paper draws out what an adopting state stands to gain or lose by engaging with the PRUs. The work speaks to ongoing debates on cross-national AI governance, dual-use technology adoption, and the operationalisation of ethical principles in real institutional settings.