The Trump administration is accelerating efforts to integrate artificial intelligence into U.S. military operations, even as some technology firms and uniformed leaders call for limits on the fast-evolving tools. Adm. Frank Bradley, who leads U.S. Special Operations Command, recently told a special forces conference in Tampa that forces must proceed carefully when applying AI to lethal decisions. He noted that while AI might one day select targets, humans must retain confidence that force will be used only as intended. Bradley oversees units responsible for the military’s most sensitive missions. His comments come as Defense Secretary Pete Hegseth presses for rapid adoption of AI systems that can operate without additional restrictions beyond existing law. Hegseth has stated that the Pentagon should employ any lawful AI capability and has criticized models that impose extra limits on military use. The administration views AI as a key advantage over China and has delayed an executive order that officials feared could slow progress. Inside the military, views differ on AI’s role. Some Pentagon officials describe it as a way to identify targets faster and shorten the time to strike. Others at Special Operations Command see it mainly as a tool for handling routine tasks, freeing personnel for core duties. Acquisition leaders there emphasize that AI should reduce workload on repetitive work rather than replace human judgment. Experts note both applications are accurate: bureaucratic uses are expanding, while combat units have already employed AI to process and share intelligence more quickly and to direct artillery fire with fewer troops. A public dispute has also emerged between the Pentagon and Anthropic over contract terms. The company raised concerns about fully autonomous weapons and mass surveillance uses, leading to accusations from administration officials and the eventual designation of the firm as a supply-chain risk, ending a large defense agreement.
