The AI-Nuclear Nexus: New CNAS Report on Managing Artificial Intelligence and Nuclear Weapons amid U.S. Rivalry with China and Russia

Washington, February 13, 2025 — Following Vice President JD Vance’s speech at the Paris Artificial Intelligence Action Summit and ahead of the Munich Security Conference, today the Center for a New American Security (CNAS) released an important new report, Averting AI Armageddon: U.S.-China-Russia Rivalry at the Nexus of Nuclear Weapons and Artificial Intelligence by Jacob Stokes, Colin H. Kahl, Andrea Kendall-Taylor, and Nicholas Lokker.

The authors argue that the nuclear order among major powers has fundamentally shifted. In particular, the People’s Republic of China (PRC) is building up its nuclear arsenal to make it numerically larger and technologically more sophisticated. As a result, the bipolar nuclear order—led by the United States and Russia—has started to give way to a more volatile tripolar one. That shift is taking place concurrently with rapid advances in artificial intelligence (AI), including for military applications. Those two trends converge in what this report calls the “AI-nuclear nexus.”

The authors write that the most critical risks of the AI-nuclear nexus can be divided into three categories: nuclear command, control, and communications (NC3); broader challenges to strategic stability; and entanglement of AI-enabled conventional systems and nuclear risks. The report explores the technical aspects of each category and analyzes how each country’s nuclear AI integration could exacerbate these risks.

The report further explains that whether new applications of AI will ultimately affect risk levels depends on three factors: the role of humans, the degree to which AI systems become a single point of failure, and AI’s effects on the offense-defense balance.

The report then places theose military-technological shifts into the larger landscape of U.S., Chinese, and Russian thinking about nuclear modernization and military AI. It also describes the context and constraints for diplomacy and other policy actions to manage the resulting risks.

Finally, the authors argue that effectively managing AI-nuclear issues in relation to China and Russia to bolster both deterrence and stability will require a proactive approach that accounts for diplomatic, military, and technological factors. Policymakers in the United States should work to better understand and address these issues in advance of the associated technologies coming into widespread use.

The authors recommend that policymakers:

• Build knowledge and competency around issues at the AI-nuclear nexus and how the United States, China, and Russia approach them;
• Integrate AI into diplomatic initiatives related to nuclear and other strategic risks, and vice versa;
• Establish and promote norms for the safe use of AI in relation to nuclear arsenals and other strategic capabilities;
• Develop policy and technical criteria for assessing exactly how and when to keep humans in the loop on all nuclear-related processes;
• Include AI technologies as a factor in oversight and reviews of the U.S. nuclear arsenal;
• Invest in AI-enabled cyber and space capabilities to enhance defense and resilience, reduce incentives to attack those areas, and mitigate entanglement risks;
• Consult closely with U.S. allies about how AI will shape extended deterrence calculations related to both nuclear and conventional capabilities; and
• Pursue a comprehensive set of risk reduction and crisis management mechanisms with China and Russia while recognizing the obstacles to progress.

The authors of the report joined the Brussels Sprouts podcast to discuss the report’s findings: listen here.

For more information or to arrange an interview with the report’s authors, please contact Alexa Whaley at [email protected].