The Trump administration is accelerating efforts to integrate artificial intelligence into the U.S. military, arguing that the technology could provide a major strategic advantage. At the same time, some technology firms and senior military leaders are urging caution, warning that powerful AI systems must be governed carefully before they are entrusted with life-and-death decisions on the battlefield.
Adm. Frank Bradley, commander of U.S. Special Operations Command, recently addressed a special operations conference in Tampa, Florida, where he stressed the importance of approaching AI with care when it comes to military applications. Bradley 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.”
Bradley said he can envision a future in which AI plays a role in identifying military targets. However, he emphasized that human oversight must remain central. According to Bradley, “we, as humans, have to have the confidence that … it’s going to deliver violence only where we intend it to be delivered.”
His comments stand in contrast to the aggressive push by Defense Secretary Pete Hegseth to expand AI capabilities across the armed forces. The debate has created tension between Pentagon leaders and some technology companies that want stronger safeguards governing how their systems can be used.
Hegseth has repeatedly argued that the military should be free to employ AI in any lawful manner necessary for national defense. Speaking to SpaceX employees in January, he declared that he would reject AI systems “that won’t allow you to fight wars” and said he favors platforms that operate “without ideological constraints that limit lawful military applications.”
The administration’s support for military AI mirrors its broader strategy of expanding American technological leadership while competing with foreign rivals, particularly China, even as concerns grow over regulation and oversight.
President Donald Trump recently scrapped plans for a new executive order on artificial intelligence just hours before it was expected to be signed at the White House. According to Trump, the proposal risked slowing America’s progress in the rapidly evolving field.
“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.
Asked about Bradley’s concerns, a Pentagon official said the department’s focus remains on developing practical battlefield tools that allow troops to identify and analyze targets more rapidly, ultimately accelerating military operations. The official spoke anonymously in order to discuss the issue candidly.
Officials within U.S. Special Operations Command have generally framed AI as a force multiplier rather than a replacement for human decision-making, describing it as a technology that allows service members to spend more time concentrating on mission execution.
Sgt. Maj. Andrew Krogman, the command’s senior enlisted leader, said he believes AI can take over routine administrative functions, freeing operators for more critical tasks while improving organizational efficiency.
Melissa Johnson, the command’s top acquisition official, similarly argued that AI should be used to lighten the burden of repetitive work.
“We’re leveraging AI more and more, but it’s not to replace operator judgment, it’s to enhance it,” she added.
Helen Toner, interim executive director of Georgetown University’s Center for Security and Emerging Technology, said both perspectives on military AI are valid.
“There are a huge number of potential uses for AI in these kinds of bureaucratic settings, which the U.S. military is actively exploring,” Toner said.
Evidence of that expansion is already visible. Lt. Gen. Michael Conley, who leads Air Force Special Operations Command, told lawmakers in May that AI-powered “bots” were used during the Iran war to rapidly downgrade highly classified intelligence into a format that could be shared with drone operators in the field within seconds.
At the same time, officials acknowledge that AI is increasingly being used not only for administrative functions but also for identifying and engaging military targets.
Toner’s research center highlighted that reality in a study released two years ago examining how the Army’s 18th Airborne Corps used artificial intelligence to guide artillery targeting. The study concluded that the corps performed “just as efficiently as the best unit in recent American history” while utilizing roughly 2,000 fewer personnel.
“Human operators are still the ones making crucial decisions, but AI … is making it possible to operate with a new level of speed and scale,” she said.
The battle over military AI, who controls it, and how it should be regulated has become unusually public during the Trump administration.
One of the most visible disputes involves Defense Secretary Hegseth and Anthropic, which has raised concerns about unrestricted government use of advanced AI systems. The company has expressed worries about technologies such as fully autonomous armed drones and large-scale AI-powered surveillance programs capable of monitoring dissent.
The conflict intensified after Anthropic CEO Dario Amodei refused to retreat from his concerns regarding the use of the company’s Claude chatbot within classified Pentagon environments. Trump and Hegseth subsequently accused the company of undermining national security.
The Pentagon later designated Anthropic, based in San Francisco, as a supply-chain risk, terminating a $200 million defense contract and barring other government contractors from partnering with the company.
Anthropic responded with a lawsuit, arguing that the government was unlawfully retaliating against the company by applying a designation intended to address threats from foreign adversaries and national security sabotage.
Since then, Pentagon officials have highlighted partnerships with Anthropic’s competitors, including Google, OpenAI, and SpaceX, as they seek AI systems capable of helping to “augment warfighter decision-making in complex operational environments.”
Toner, who previously served on OpenAI’s board before a high-profile dispute with CEO Sam Altman, said many Americans underestimate how cautiously military leaders typically approach emerging 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.
{Matzav.com}