From Divergence to Cooperation

The adoption of AI as a force multiplier in military contexts (“military AI”) has spurred a new era of defense innovation. In July 2025, the United States unveiled its AI Action Plan, a strategic roadmap for securing leadership in the global AI race that includes a vision for integrating AI into defense. The plan asserts that “AI has the potential to transform both the warfighting and back-office operations of the Department of Defense (DoD)” and that, to maintain military superiority, the US Armed Forces must “aggressively adopt” AI.

The plan proposes to build robust AI infrastructure for defense and national security. Recommendations include building high-security data centers for defense and intelligence use, establishing a virtual proving ground for testing AI-enabled defense technologies, and securing strategic agreements with cloud-service providers and other private sector entities to ensure operational resilience during national emergencies. To develop and sustain these capabilities, the plan emphasizes building a skilled AI workforce within the DoD by identifying critical skill sets and establishing talent development programs. The plan foresees growing “Senior Military Colleges into hubs of AI research, development, and talent building”, with AI-specific curricula “including in AI use, development, and infrastructure management”.

Across the Atlantic, the European AI Continent Action Plan (2024) similarly identifies foundational technologies such as AI as critical to long-term economic and military strength. This emphasis is echoed in the Rearm Europe White Paper, which mentions a forthcoming European Armament Technological Roadmap. That roadmap will aim at “leveraging investment into dual-use advanced technological capabilities at EU, national, and private level[s]” with an initial focus on AI and Quantum.

Momentum on both sides of the Atlantic aligns with NATO’s drive to harness AI for collective security through initiatives that integrate AI across military operations, innovation, education, and capability development. Yet, notable differences may complicate coordination within NATO. Unlike the US AI Action Plan, the European strategy documents mentioned above lack specific recommendations pertaining to AI in defense, revealing a possible gap in ambition and implementation. Defense also remains the prerogative of individual EU member states, which potentially hinders a unified Europe-wide approach. Responsible deployment poses another challenge. While the US plan calls for systems to remain “secure and reliable”, it stops short of defining these terms. Brussels, by contrast, has issued a white paper on developing responsible, ethical, and trustworthy AI systems for defense. The paper references NATO’s Principles of Responsible Use, which prioritize lawfulness, responsibility and accountability, explainability and traceability, reliability, governability, and bias mitigation, among other international standards. Addressing these divergences will be key to enabling effective and sustained transatlantic cooperation on military AI.

The author acknowledges Stephen Silvestri, GMF Tech summer trainee, for his research support on this piece.