The nuclear-weapons states China, France, Russia, the United Kingdom, and the United States are increasingly recognising the implications of integrating AI into nuclear weapons command, control, and communication systems. Exploring the risks inherent to today’s advanced AI systems, this report sheds light on characteristics and risks across different branches of this technology and establishes the basis for a general purpose risk assessment framework.
Supported by the Bureau of Arms Control, Verification and Compliance (AVC), this ELN project examines the potential impact of AI-enabled decision-making tools on strategic decision-making, particularly in crisis scenarios, within the context of the nuclear domain. This exploration encompasses an analysis of the benefits and challenges presented by AI in this realm.
The project’s core objective is to evaluate AI-associated risks, with a specific emphasis on the incorporation of AI-enabled tools into nuclear command, control, and communications (NC3) systems, and how it influences the process of nuclear decision-making. By reviewing open-source literature from British, Chinese, French, and Russian sources, in conjunction with insights gathered from two workshops, it will identify areas of consensus, potential obstacles, and practical policy recommendations aimed at mitigating risks among the P5 states.
Chinese thinking on AI integration and interaction with nuclear command and control, force structure, and decision-making
Fei Su and Jingdong Yuan analyse Chinese-language literature to present Chinese perspectives on AI and its military applications. The paper offers recommendations to mitigate the risks associated with the military use of AI in nuclear C2 systems, particularly focusing on the steps that China could consider to enhance its practices.
Russian thinking on AI integration and interaction with nuclear command and control, force structure, and decision-making
Oleg Shakirov analyses Russian-language literature on the Russian debate on AI and the nuclear field and offers recommendations for P5 states to advance dialogue on AI integration into nuclear C2, force structure and decision-making.
UK thinking on AI integration and interaction with nuclear command and control, force structure, and decision-making
Alice Saltini analyses the British literature on the UK’s perception of military and nuclear applications of AI and their impact on strategic stability and NC3. The paper offers recommendations for unilateral measures that the UK can take, as well as multilateral initiatives within the P5 framework, to address the risks associated with AI in nuclear decision-making
French thinking on AI integration and interaction with nuclear command and control, force structure, and decision-making
Héloïse Fayet analyses the French literature on France’s perception of military AI, especially its consequences on strategic systems and competition, and nuclear deterrence. Fayet offers practical recommendations for France both domestically and internationally.