solution · S05

International AI Arms Control Treaties

Binding international agreements to regulate the development and deployment of AI in military applications, modeled on nuclear non-proliferation and chemical weapons conventions.

Updated

Observatory assessment: 18 / 100. Policy Debate

How assessments are made

Deep dive

The absence of international norms governing military AI creates a dangerous vacuum, now characterized by a three-way tension between states, corporate leadership, and the technical workforce. While intergovernmental forums like the UN CCW see slow progress, a significant new front has opened with widespread advocacy from within the AI community itself. A recent letter from over a thousand AI researchers and engineers warned of a catastrophic arms race, while a separate movement of AI lab employees has called for an international slowdown with verifiable enforcement. This bottom-up pressure complements the top-down dynamic of corporate ethical red lines, such as Anthropic's 2026 refusal to grant the Pentagon unrestricted use of its AI for lethal autonomous weapons. While states continue to pursue national security interests, evidenced by rising US-China tensions and unilateral control considerations, they now face a complex landscape where both corporate policy and employee activism act as powerful, parallel forms of private governance in the military AI supply chain.

Trajectory and projections

Key players

  • United Nations (CCW framework)
  • Campaign to Stop Killer Robots (civil society)
  • ICRC (International Committee of the Red Cross)
  • US State Department (Political Declaration on Military AI)
  • AUKUS & NATO (allied AI governance frameworks)
  • Major AI Labs (e.g., Anthropic, OpenAI)
  • ControlAI (non-profit advocacy)
  • AI Researchers & Engineers (advocacy groups)
  • Academic Institutions (e.g., Stanford HAI)
  • Philanthropic Funds (e.g., EA Funds)

Barriers

  • Major military powers (US, China, Russia) resist binding limits
  • Verification challenge: AI capabilities are software-based and hard to inspect
  • Dual-use technology: civilian AI easily adapted for military purposes
  • Speed of development outpaces diplomatic timelines
  • Conflict between corporate AI ethics and national security demands

Top evidence

The Economics of AI Counterproliferation

RAND Corporation

This article explores the economic aspects of AI counterproliferation, focusing on strategies and challenges related to controlling the spread of advanced AI capabilities, particularly those with potential for military misuse.

UK-Germany joint statement on advanced AI safety and security

UK DSIT AI (GOV.UK)

The UK and Germany have issued a joint statement on advanced AI safety and security, indicating international cooperation and policy efforts to manage the risks associated with frontier AI systems. This highlights a commitment to developing shared standards and approaches for AI governance.

Serial No. 119-42 - DEEPSEEK AND UNITREE ROBOTICS: EXAMINING THE NATIONAL SECURITY RISKS OF PRC ARTIFICIAL INTELLIGENCE, ROBOTICS, AND AUTONOMOUS TECHNOLOGIES AND BUILDING A SECURE U.S. TECHNOLOGY BASE

GovInfo Congressional Hearings

This congressional hearing examines the national security risks posed by Chinese artificial intelligence, robotics, and autonomous technologies, specifically mentioning DeepSeek and Unitree Robotics. It also aims to discuss strategies for building a secure U.S. technology base in response to these threats.

A Red Line and Oversight Framework for Government AI Contracts

Alignment Forum

This article proposes a framework for establishing 'red lines' and oversight in government AI contracts, aiming to prevent harmful uses like autonomous targeting without human control and untargeted profiling. It focuses on mechanism design to ensure responsible AI deployment while allowing beneficial applications.

New Stanford Grants Tackle AI's Impact on Global Security and Geopolitics - Stanford HAI

Stanford HAI

Stanford University has awarded new grants to research the multifaceted impact of AI on global security and geopolitics. This initiative aims to deepen understanding of how AI technologies influence international relations, conflict, and stability.