The Biden‑era Trump administration has declared that artificial intelligence (AI) is the next winning edge that will allow the United States to outpace rival powers in unknown future conflicts. At a recent annual special‑forces conference in Tampa, Adm. Frank Bradley, the commander of U.S. Special Operations Command, urged that “we have to be very careful about how we [embrace] AI’s employment and the impact on the delivery of lethality.” He warned that if AI systems are allowed to pick targets, commanders must retain confidence that the system will only deploy violence where the human operator intends it.
That caution co‑exists with Defense Secretary Pete Hegseth’s rapid‑deployment push for AI‑based “advances.” Hegseth told a SpaceX auditorium in January that the Pentagon would reject any AI model that “won’t allow you to fight wars.” He insisted that systems must operate without “ideological constraints that limit lawful military applications.”
The two positions illustrate the divide within the armed forces. A Pentagon official, speaking under a pseudonym, said the goal is to create “functional battlefield tools” that pave the way for quicker target identification and faster strike capabilities. In contrast, Sgt. Maj. Andrew Krogman of Special Operations said AI will be used mainly for administrative tasks to free up operators, while acquisition chief Melissa Johnson added that AI should “reduce the cognitive workload on mundane tasks.” 2026 Air Force Special Operations Command commander Lt. Gen. Michael Conley revealed that the unit had used “bot”‑style AI to convert top‑secret intelligence into secret‑level data in seconds, enabling faster drone‑pilot coordination.
These use cases are complemented and challenged by new research from Georgetown’s Center for Security and Emerging Technology, where Helen Toner reports on how the Army’s 18th Airborne Corps used AI for artillery targeting with speed comparable to a legendary unit but with 2,000 fewer personnel.
However, AI is also helping the military strike targets. In interviews the Pentagon has said contracting companies such as Anthropic are being scrutinized for concerns about fully autonomous drones, mass surveillance, and potential misuse of the technology. In a bitter contract dispute, Anthropic’s CEO Dario Amodei defied Pentagon demands to adjust the company’s chatbot Claude for classified use, prompting Hegseth and Trump to declare the firm was a “supply‑chain risk.” The Pentagon subsequently canceled a $200 million defense contract and prevented other contractors from collaborating with Anthropic.
Anthropic filed a lawsuit accusing the Pentagon of illegal re‑branding the company as a sabotage threat. In the same period, the Pentagon announced it would turn to rivals such as Google, OpenAI and SpaceX for AI solutions to “augment war‑fighter decision‑making.” Owens‑Edge voices say the Pacific‑wide discourse revolves around balancing swift operational advantage with the moral obligation to limit unintended consequences like friendly fire or civilian harm.
Meanwhile, President Donald Trump has repeatedly signaled concern that any move to formalize AI policy will undermine the United States’ lead over China. Trump called off a planned AI executive order just hours before a White House ceremony – declaring that “we’re leading China…and I don’t want to do anything that’s going to get in the way of that lead.”
The US AI‑military dialogue therefore sits at a crossroads: a push to keep the country technologically ahead versus an emerging consensus that any advantage must come with a robust safety framework, accountability, and public scrutiny. As the Pentagon pursues partnerships with the world's most advanced AI firms, tech leaders are demanding checks and balances to prevent the creation of an unchecked autonomous weapon system that could jeopardize national security and civil liberties alike. The debate underscores a moral imperative: that the innovations necessary for national‑security benefits can coexist with a governance structure that protects both US interests and global stability.
That caution co‑exists with Defense Secretary Pete Hegseth’s rapid‑deployment push for AI‑based “advances.” Hegseth told a SpaceX auditorium in January that the Pentagon would reject any AI model that “won’t allow you to fight wars.” He insisted that systems must operate without “ideological constraints that limit lawful military applications.”
The two positions illustrate the divide within the armed forces. A Pentagon official, speaking under a pseudonym, said the goal is to create “functional battlefield tools” that pave the way for quicker target identification and faster strike capabilities. In contrast, Sgt. Maj. Andrew Krogman of Special Operations said AI will be used mainly for administrative tasks to free up operators, while acquisition chief Melissa Johnson added that AI should “reduce the cognitive workload on mundane tasks.” 2026 Air Force Special Operations Command commander Lt. Gen. Michael Conley revealed that the unit had used “bot”‑style AI to convert top‑secret intelligence into secret‑level data in seconds, enabling faster drone‑pilot coordination.
These use cases are complemented and challenged by new research from Georgetown’s Center for Security and Emerging Technology, where Helen Toner reports on how the Army’s 18th Airborne Corps used AI for artillery targeting with speed comparable to a legendary unit but with 2,000 fewer personnel.
However, AI is also helping the military strike targets. In interviews the Pentagon has said contracting companies such as Anthropic are being scrutinized for concerns about fully autonomous drones, mass surveillance, and potential misuse of the technology. In a bitter contract dispute, Anthropic’s CEO Dario Amodei defied Pentagon demands to adjust the company’s chatbot Claude for classified use, prompting Hegseth and Trump to declare the firm was a “supply‑chain risk.” The Pentagon subsequently canceled a $200 million defense contract and prevented other contractors from collaborating with Anthropic.
Anthropic filed a lawsuit accusing the Pentagon of illegal re‑branding the company as a sabotage threat. In the same period, the Pentagon announced it would turn to rivals such as Google, OpenAI and SpaceX for AI solutions to “augment war‑fighter decision‑making.” Owens‑Edge voices say the Pacific‑wide discourse revolves around balancing swift operational advantage with the moral obligation to limit unintended consequences like friendly fire or civilian harm.
Meanwhile, President Donald Trump has repeatedly signaled concern that any move to formalize AI policy will undermine the United States’ lead over China. Trump called off a planned AI executive order just hours before a White House ceremony – declaring that “we’re leading China…and I don’t want to do anything that’s going to get in the way of that lead.”
The US AI‑military dialogue therefore sits at a crossroads: a push to keep the country technologically ahead versus an emerging consensus that any advantage must come with a robust safety framework, accountability, and public scrutiny. As the Pentagon pursues partnerships with the world's most advanced AI firms, tech leaders are demanding checks and balances to prevent the creation of an unchecked autonomous weapon system that could jeopardize national security and civil liberties alike. The debate underscores a moral imperative: that the innovations necessary for national‑security benefits can coexist with a governance structure that protects both US interests and global stability.
