The Trump administration is looking to push artificial intelligence in the U.S. military as it faces pressure from military leaders and companies to implement safeguards.

In a recent annual special forces conference in Tampa, Florida, Adm. Frank Bradley, head of U.S. Special Operations Command, told attendees that troops “have to be very careful about how we come to (AI’s) employment and its inspiration into the delivery of lethality.” 

“We, as humans, have to have the confidence that … it’s going to deliver violence only where we intend it to be delivered,” Bradley said. 

In response to the remarks, a Pentagon official, speaking on the condition of anonymity, told the Monterey Herald, said the Pentagon was focussing on efforts to make “functional battlefield tools” with AI to help troops to identify targets quickly. 

Meanwhile, U.S. Special Operations Command officials said AI should not be a tool for eradicating targets, but to assist troops to focus on their mission. Sgt. Maj. Andrew Krogman said at the conference that AI could be used for administrative tasks or to modernize workflows. 

AI has already been used by the military to identify targets, including the use of Palantir's Maven Smart System, which integrates Anthropic’s Claude AI to speed up targeting and planning. 

Defense Secretary Pete Hegseth is pushing to integrate AI into the military. 

Hegseth said he would reject AI models that “won’t allow you to fight wars,” in a recent speech at Elon Musk’s space flight company, SpaceX in south Texas. He also said that military AI systems should operate “without ideological constraints that limit lawful military applications.” 

President Trump has also championed AI development as an economic engine, and pledged to undo AI safety regulations from the Biden era. 

However, the administration’s push to integrate military AI has been complicated by its legal battle with Anthropic. On Feb. 27, Trump ordered U.S. agencies to stop using Anthropic's AI models after CEO Dario Amodei declined the Pentagon’s demands to allow unrestricted use of its AI model, citing dangers around autonomous drones and mass surveillance. 

Meanwhile, Republicans are divided over whether to embrace the AI industry or to approach it with caution, as voters express fears over its impact on American society, jobs, and energy costs. 

Concerns have also grown within the Trump administration. At the Treasury Department’s headquarters, Secretary Scott Bessent and former Federal Reserve Chair Jerome Powell convened an urgent meeting warning against cybersecurity risks posed by Anthropic’s advanced Claude Mythos model. 

Nevertheless, President Trump on May 21 withdrew from a plan to sign a new AI executive order hours before an expected White House ceremony due to concerns that it could soften America’s tech edge. The order was aimed at developing a framework for the government to vet national security risks from new AI systems before releasing them publicly, according to AP News.

“We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead,” Trump told reporters. 

The Pentagon has reached deals with Anthropic competitors, including Google, OpenAI and Space X, for AI technology that will “augment warfighter decision-making in complex operational environments,” the Defense Department said.

“The president wants us to be pro-innovation. He wants us to win the AI race against all other countries in the world,” said Vice President JD Vance said Tuesday at a White House press briefing. "We also want to make sure that we’re protecting people.” 

When asked about security risks, he said the administration was taking a collaborative approach with tech companies.