The Trump administration is accelerating the integration of artificial intelligence into military operations while facing resistance from technology companies and cautionary voices from military leaders about the need for safety measures.

Defense Secretary Pete Hegseth has pushed for rapid AI adoption, telling SpaceX employees in January that he would reject AI models "that won't allow you to fight wars" and advocating for systems operating "without ideological constraints that limit lawful military applications." The Pentagon has emphasized developing AI technology that can "augment warfighter decision-making in complex operational environments."

However, senior military officials have expressed more measured approaches. Admiral Frank Bradley, head of U.S. Special Operations Command, told a recent conference that troops "have to be very careful about how we come to AI's employment." While Bradley said he could envision AI determining targets, he emphasized that "we, as humans, have to have the confidence that it's going to deliver violence only where we intend it to be delivered."

The debate has created public disputes between the Pentagon and AI companies. Anthropic, maker of the Claude chatbot, has clashed with the Defense Department over concerns about unchecked government use of its technology, including fully autonomous weapons and AI-assisted mass surveillance. After Anthropic CEO Dario Amodei refused to back down on safety concerns, the Pentagon labeled the company a supply chain risk, ending a $200 million contract and prohibiting other contractors from working with the firm.

Anthropic has sued the Pentagon, claiming illegal retaliation through a designation meant to protect against foreign adversaries. The company cited concerns about autonomous armed drones and surveillance systems that could track dissent.

Military AI applications currently range from administrative tasks to target identification. During recent operations, Air Force Special Operations used AI to declassify intelligence within seconds for drone operators. A 2022 study found the Army's 18th Airborne Corps used AI to target artillery strikes as efficiently as top-performing units while using 2,000 fewer personnel, though human operators maintained final decision-making authority.