The Trump administration is pushing to use artificial intelligence in the U.S. military even as it faces calls for caution from some companies and military leaders. Adm. Frank Bradley, head of U.S. Special Operations Command, told attendees of a recent special forces conference in Tampa, Florida, that troops “have to be very careful about how we come to (AI’s) employment and its inspiration into the delivery of lethality.” Bradley envisions a future where AI determines targets but insists that humans must have confidence that violence is delivered only where intended.
Diverging Views on AI's Role
Bradley’s remarks contrast with Defense Secretary Pete Hegseth’s push to rapidly evolve the military through AI, leading to clashes with tech companies over safety. Hegseth has insisted the Pentagon can use AI in any legal way, rejecting models that “won’t allow you to fight wars.” President Donald Trump canceled plans for a new AI executive order, fearing it could dull America’s edge. A Pentagon official, speaking anonymously, said efforts focus on creating “functional battlefield tools” to speed up target identification and strikes.
At U.S. Special Operations Command, officials view AI as a tool to handle administrative tasks, freeing troops for missions. Sgt. Maj. Andrew Krogman sees AI modernizing operations, while Melissa Johnson, top acquisition official, says AI should “reduce cognitive workload on mundane tasks” without replacing operator judgment. Helen Toner of Georgetown University notes both descriptions are true, highlighting AI’s potential in bureaucratic settings and targeting.
AI in Action: Targeting and Intelligence
Lt. Gen. Michael Conley told Congress that during the Iran war, AI “bots” converted top secret intelligence to secret classification in seconds for drone operators. The Army’s 18th Airborne Corps used AI to target artillery strikes as efficiently as the best historical unit but with 2,000 fewer personnel. Toner emphasizes that human operators still make crucial decisions, but AI enables new speed and scale.
Public Dispute with Anthropic
The clash over AI integration has played out publicly. Hegseth and Anthropic are in a contract dispute over concerns about unchecked government use, including autonomous armed drones and mass surveillance. After CEO Dario Amodei refused to back down, the Pentagon labeled Anthropic a supply chain risk, ending a $200 million contract and prohibiting other contractors from working with the company. Anthropic sued, alleging illegal retaliation. The Pentagon now turns to rivals like Google, OpenAI, and SpaceX for AI that can “augment warfighter decision-making.”
Toner, a former OpenAI board member, says the public often underestimates the military’s caution with new technologies. “Commanders want their missions to succeed, which means both being able to create lethal effects at scale, and avoiding unintended effects like friendly fire, civilian casualties, or simply identifying targets incorrectly,” she said.
