It draws on official strategies, doctrines, and speeches, reports and studies written by French scholars, and articles written by former or active military officers during their higher military education.

The paper gives a brief overview of the French approach to the debate, followed by the official positions on the development of military AI. It then focuses on the impact of AI on C2 and decision systems, including NC3 and its consequences on strategic stability. It concludes with a short overview of how France understands military AI programs from the P5 countries.

The paper recommends that, both on a domestic and international level, France should:

• Build a stronger community of researchers, officials and the private sector on AI and its impact on nuclear deterrence. Even though some initiatives such as the Réseau Nucléaire & Stratégie aim at creating a new generation of strategists and researchers on nuclear policy and industry, it could be interesting to fund additional scholarships, to create a chair on these topics in academia, or to support a seminar on this subject.
• Strengthen the links between the public and the private sectors. Misunderstanding about AI can fuel wrong analyses about its impact on military and nuclear strategy. Getting together data scientists and computer scientists with IR specialists and nuclear experts could be helpful.
• Support a P5 initiative on AI and risks related to the NC3. The P5 process on Strategic Risk Reduction, including the Youth Group, could be a good arena to push for these topics. A statement about the necessity to keep humans in the loop and to introduce strong means of crisis communication in order to avoid inadvertent escalation could be a good initiative.