America’s military should not fear open-source AI. It should harness it.

ABOUT THE AUTHOR: Patrick McSpadden is a retired U.S. Air Force lieutenant colonel and former intelligence officer with more than 21 years of service, including multiple operational deployments. He writes on defense policy, military technology and national security. The views expressed are his own and do not reflect the official position of the Department of Defense or the U.S. government.

Washington is asking the wrong question about artificial intelligence. The debate has become whether America’s most capable AI models should be open or closed. For the Department of Defense, the more important question is simpler: which approach gives American warfighters the greatest operational advantage?

Open-source AI is more than a catalyst for commercial innovation. It is a strategic advantage for the U.S. military, enabling American companies to push the technology forward while giving the DOD a foundation it can inspect, secure, adapt and deploy faster than any adversary.

History is full of militaries that possessed superior technology but failed to adapt it before their adversaries. America has avoided that trap because its strength has never rested solely on government laboratories. It has rested on a dynamic private sector capable of solving problems at remarkable speed and scale. Artificial intelligence should follow that same model.

The United States does not gain its greatest advantage from a single frontier model. It gains it from thousands of companies, universities, software developers, defense contractors and startups building on top of those models. Open-source AI dramatically expands that talent pool.

When capable open models are available, innovation accelerates. Small defense startups can develop mission-specific applications without spending billions to build foundation models from scratch. Universities can improve efficiency and security. Established defense firms can focus on operational capabilities instead of duplicating basic research. Air Force and Space Force software factories can rapidly prototype tools for operators rather than waiting for commercial vendors to prioritize military requirements.

That innovation eventually finds its way into uniform.

The Pentagon does not need AI that wins internet benchmarks. It needs AI that helps airmen generate tomorrow’s air tasking order, intelligence analysts connect fragmented reporting, maintainers return aircraft to the fight faster, logisticians anticipate supply shortfalls, and commanders make better decisions under pressure. Those are operational problems, not consumer ones.

Open models offer unique advantages in solving them.

Unlike purely proprietary systems, open models can be inspected, modified and fine-tuned for specific missions. They can be deployed inside government-controlled environments, optimized for disconnected operations, and integrated into classified networks under appropriate security controls. That flexibility allows military organizations to shape AI around operational requirements instead of adapting operations to commercial software.

None of this suggests every military application should rely on open models. Certain missions involving highly sensitive intelligence sources, advanced weapons systems, or strategic planning will continue to require tightly controlled proprietary capabilities operating inside the government’s most secure environments. The Department of Defense does not need an ideological commitment to either approach. It needs the judgment to determine where closed models provide the greatest security advantage and where open models deliver greater speed, adaptability and operational value.

That distinction has become increasingly important as leading AI companies debate the future of open-source development.

Anthropic has been among the most prominent voices urging caution over open models, arguing that broad public release could increase security risks. But the U.S. has decades of experience managing risk while embracing and proliferating open-source technologies. And Anthropic may prefer tighter limits on open-source AI because those limits ultimately reduce competition for its own systems, keep more control in its hands, and allow it to frame that stance as “safety-first” rather than self-interest. But DOD leaders are not deciding how to protect a proprietary product. They are deciding how to maintain America’s military advantage. And America has rarely won strategic competitions by narrowing participation in innovation. It has won by expanding it.

The internet, GPS, cloud computing and modern software all evolved through collaboration among government, industry, academia and entrepreneurs. The military benefited because it could adapt technologies already advancing at commercial speed rather than developing every capability itself.

Artificial intelligence presents the same opportunity, if the U.S. can ensure its own innovators continue moving faster than China’s.

The United States must also ensure its own innovators continue moving faster than anyone else.

Every defense startup improving logistics software, every university refining model efficiency, every software factory developing mission applications, and every commercial company building new AI tools contributes to a competitive advantage that no centralized system can easily replicate.

America does not need to choose between security and innovation, or between open and closed AI. It needs the discipline to employ each where it offers the greatest operational advantage.

America’s AI advantage will not be measured solely by the sophistication of its frontier models. It will be measured by how quickly American companies can innovate, how effectively the Department of Defense can translate those innovations into operational capability, and how consistently U.S. warfighters stay one step ahead of their adversaries. That is the real promise of open-source AI. Not replacing closed models, but ensuring America’s greatest strategic advantage remains the speed with which we turn innovation into military power.