The Trump administration is actively scaling the integration of artificial intelligence into U.S. military operations, sparking internal debate among defense leadership regarding safety and control. While Defense Secretary Pete Hegseth advocates for systems free from ideological constraints to ensure battlefield superiority, some high-ranking officers emphasize the necessity of maintaining human oversight. Adm. Frank Bradley, head of U.S. Special Operations Command, stressed at a recent conference in Tampa, Florida, that commanders must maintain confidence in the delivery of lethal force, ensuring AI acts only as intended.

The Pentagon’s AI application currently spans two distinct areas: bureaucratic efficiency and direct combat targeting. Officials from the U.S. Special Operations Command described using AI primarily to reduce the cognitive workload of administrative tasks and to optimize mission planning. Lt. Gen. Michael Conley noted that Air Force units successfully employed AI to convert classified intelligence for rapid sharing with drone operators during the Iran war. Simultaneously, academic experts highlight that AI has already demonstrated the ability to execute artillery targeting with efficiency comparable to elite units while requiring 2,000 fewer service members.

This strategic push has led to public friction between the Department of Defense and AI companies over safety guardrails. A notable contract dispute between the Pentagon and Anthropic resulted in the company being labeled a supply chain risk after CEO Dario Amodei expressed concerns regarding the government’s use of its technology in autonomous drones and mass surveillance. The Pentagon subsequently terminated a $200 million contract with the company and turned to alternative providers, including Google, OpenAI, and SpaceX. Anthropic has filed a lawsuit, alleging that the designation is an illegal form of retaliation. Despite these tensions, defense policy experts maintain that military commanders remain focused on balancing the need for speed and scale with the risks of unintended outcomes, such as civilian casualties or friendly fire.

The Trump administration is actively scaling the integration of artificial intelligence into U.S. military operations, sparking internal debate among defense leadership regarding safety and control. While Defense Secretary Pete Hegseth advocates for systems free from ideological constraints to ensure battlefield superiority, some high-ranking officers emphasize the necessity of maintaining human oversight. Adm. Frank Bradley, head of U.S. Special Operations Command, stressed at a recent conference in Tampa, Florida, that commanders must maintain confidence in the delivery of lethal force, ensuring AI acts only as intended.

The Pentagon’s AI application currently spans two distinct areas: bureaucratic efficiency and direct combat targeting. Officials from the U.S. Special Operations Command described using AI primarily to reduce the cognitive workload of administrative tasks and to optimize mission planning. Lt. Gen. Michael Conley noted that Air Force units successfully employed AI to convert classified intelligence for rapid sharing with drone operators during the Iran war. Simultaneously, academic experts highlight that AI has already demonstrated the ability to execute artillery targeting with efficiency comparable to elite units while requiring 2,000 fewer service members.

This strategic push has led to public friction between the Department of Defense and AI companies over safety guardrails. A notable contract dispute between the Pentagon and Anthropic resulted in the company being labeled a supply chain risk after CEO Dario Amodei expressed concerns regarding the government’s use of its technology in autonomous drones and mass surveillance. The Pentagon subsequently terminated a $200 million contract with the company and turned to alternative providers, including Google, OpenAI, and SpaceX. Anthropic has filed a lawsuit, alleging that the designation is an illegal form of retaliation. Despite these tensions, defense policy experts maintain that military commanders remain focused on balancing the need for speed and scale with the risks of unintended outcomes, such as civilian casualties or friendly fire.