Think tank · Policy · Pennsylvania Ave NW, Washington · Est. 2007
CNAS AI Security & Stability is a CNAS (Center for a New American Security) project that aims to inform U.S. government decision-making on AI policy issues tied to national security and strategic stability, with an emphasis on risks and policy levers arising from the U.S.-China AI competition. The project is described as a cross-program effort spanning CNAS’s Technology and National Security, Defense, Indo-Pacific Security, and Energy, Economics, and Security programs, with work organized around five major lines of effort: (1) mitigating national-security-relevant risks from advanced AI capabilities; (2) shaping “compute governance” policies; (3) improving U.S. military processes for safe, secure, and trusted AI; (4) analyzing Chinese decision-making and its implications for stability; and (5) informing how U.S. economic security tools can shape AI development and proliferation.
No curated programs yet. Check the careers page ↗.
Developing strong, pragmatic and principled national security and defense policies.
The Axis of Upheaval Goes to WarDeveloping strong, pragmatic and principled national security and defense policies.
Sharper: Countering Cyber ThreatsDeveloping strong, pragmatic and principled national security and defense policies.
How to Avoid Disaster in the Next Iran WarDeveloping strong, pragmatic and principled national security and defense policies.
How Much More Trump Trouble Can Indonesia’s Prabowo Take?Developing strong, pragmatic and principled national security and defense policies.
Sea Drones a Taiwan Game ChangerDeveloping strong, pragmatic and principled national security and defense policies.
The Trump-Xi Summit Leaves Taiwan in LimboDeveloping strong, pragmatic and principled national security and defense policies.
SharperDeveloping strong, pragmatic and principled national security and defense policies.
Mr. Xi Goes to Washington: An Opportunity for U.S.-China AI AssuranceDeveloping strong, pragmatic and principled national security and defense policies.
Japan’s New CIA Could Be Takaichi’s Trump CardDeveloping strong, pragmatic and principled national security and defense policies.
A $5,000 Bike Shows Why It’s Hard to Build in AmericaDeveloping strong, pragmatic and principled national security and defense policies.
Americas Secret Weapon In The Nuclear RaceExternal reporting discussed CNAS’s work and interest in advance-AI risk mitigation themes, including secure governance concepts related to advanced systems.
CNAS Insights | Washington Can’t Afford to Ignore AI’s Warning ShotCNAS published an AI policy insight authored by Ruby Scanlon and Janet Egan.
American AI Companies Can’t Get Enough ChipsCNAS published a report arguing that, in 2026, AI chip production has become a binding constraint on the pace of AI compute buildout and outlining policy implications.
CNAS Insights | MATCHing Policy to Strategy: The Bill to Control Chipmaking EquipmentCNAS published a policy analysis on the rationale for chipmaking-equipment controls as part of U.S. national security efforts related to AI.
CNAS Expands AI Policy Research with New AppointmentsCNAS announced three staff hires—Daniel Remler, James Sanders, and Liam Epstein—as it expands its multiyear Artificial Intelligence (AI) Security and Stability Initiative.
Artificial Intelligence Security and Stability project page (lines of effort and scope)CNAS’s project description page lays out the initiative’s purpose and five lines of effort for AI security and stability policy research.