---

📌 Editor’s Note: This article is part of our AI Policy Corner series, a collaboration between the Montreal AI Ethics Institute (MAIEI) and the Governance and Responsible AI Lab (GRAIL) at Purdue University. The series provides concise insights into critical AI policy developments from the local to international levels, helping our readers stay informed about the evolving landscape of AI governance. This piece covers the concerns involved with the US government’s partnerships with private AI companies as part of its security policy.

---

In July 2025, the U.S. Department of Defense’s Chief Digital and Artificial Intelligence Office (CDAO) announced contractual agreements with four frontier AI companies, namely Anthropic, Google, OpenAI and xAI, each worth up to $200 million. The same month, the White House’s AI Action Plan declared that it is ‘‘a national security imperative’’ for the United States to achieve ‘‘unquestioned and unchallenged global technological dominance’’ in AI. These developments provide a starting point for my examination of U.S. AI security policy. Drawing on government documents published between 2019 and 2026, including authorization acts, executive orders, and Department of Defense directives, I investigated how U.S. institutions frame AI in terms of strategic competition and cooperation.

When these documents discuss AI competition, they are more likely to discuss states. Nearly three-quarters of competitive framings of Artificial Intelligence concerns military advantage or rivalry with other states. When they discuss cooperation on the other hand, it mostly involves partnerships with private sectors and industry. Indeed, this finding is not entirely surprising; in practical terms, the Stanford HAI 2026 AI Index Report found that industry produced more than 90% of notable frontier AI models in 2025.

Government partnership with industry is not a new phenomenon. The U.S. government has a long history of partnering with private industry to advance technological development. Consider NASA’s extensive reliance on industrial contractors during the Apollo program, and other federally-funded research which supported much of modern computing. What is different with AI is the direction of dependence. In earlier eras, the state took the early risks and firms commercialized the results. Today, the most capable AI systems are first built by a relatively small number of companies. As a result, the government depends on industry to commercialize publicly supported innovations as well as to develop and supply some of the most capable AI systems.

This raises important concerns. The first is a question about accountability. State agreements or alliances like the North Atlantic Treaty use pacts that the Senate must approve, and executive agreements must be reported to Congress and, in most cases, published and circulated. On the other hand, partnerships with private sectors may work differently. State-private cooperation may use contracts, and flexible procurement deals whose details are often treated as confidential business information; The GAO has found that the Pentagon cannot tell whether many of its prototype agreements ever led to production or delivered capabilities to troops. Other democratic oversights may also be compromised.

The second is divided interests. Frontier AI companies answer to investors, global customers and commercial strategies of their own. Their interests may often align with U.S. security goals, but not always, and thus, how can the U.S. government ensure that the commercial decisions of private firms remain consistent with national security priorities when the government increasingly depends on those companies for critical AI capabilities? In turn, how much practical influence do private companies have over security policy?

The third concern is the potential marginalization of allies. If the United States treats companies as its primary AI partners, allied state partners and governments may be pushed to the margins on decisions that affect them. This could complicate efforts to establish shared international rules and standards for military AI. As Washington increasingly competes with states while depending on private firms, the implications extend beyond government institutions to the companies developing these technologies and the public whose security and democratic accountability are at stake.

Further Reading:

• The AI Diffusion Framework: Securing U.S. AI Leadership While Preempting Strategic Drift
• Responsible Procurement of Military AI
• Executive Order 14409

Image source: Reuters