Adm. Frank Bradley, leader of US Special Operations Command, emphasized at a recent special forces conference in Tampa, Florida, the necessity of careful consideration regarding how AI should be employed in military operations, particularly concerning its lethal applications.
Bradley envisions a future where AI identifies targets but insists that “we, as humans, must be confident that … it will only deliver violence where we intend it to be delivered.”
His comments highlight a tension between the need for safeguards and Defense Secretary Pete Hegseth’s drive to accelerate military advancements through AI. This push has prompted concerns from some technology firms regarding safety protocols.
Hegseth has advocated for the Pentagon’s unrestricted use of AI technology, asserting that he would reject any models “that won’t allow you to fight wars.” He envisions systems that function “without ideological constraints that limit lawful military applications.”
AI’s military integration is part of the Republican administration’s broader goal to enhance capabilities viewed as uniquely American, even as it grapples with pressures to implement responsible safeguards.
President Donald Trump abruptly canceled plans for a new AI executive order shortly before a scheduled signing ceremony, concerned that the move could jeopardize the US’s leading position in AI technology.
“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 stated to reporters.
In response to Bradley’s remarks, a Pentagon official clarified that efforts are currently focused on developing “functional battlefield tools” to assist troops in quickly identifying targets, thereby accelerating strikes. This official spoke anonymously to provide candid insights.
Officials from US Special Operations Command portrayed AI not as a means to eliminate targets, but as a tool that allows personnel more time to concentrate on their missions.
Sgt. Maj. Andrew Krogman, the top enlisted official at US Special Operations Command, mentioned at the conference that he foresees AI taking over administrative tasks, thereby enabling operators to focus on operational demands.
Melissa Johnson, the command’s chief acquisition official, noted that AI should work towards “reducing the cognitive workload on mundane tasks.”
“We’re leveraging AI more and more, but it’s not to replace operator judgment; it’s to enhance it,” she stated.
Helen Toner, interim executive director at Georgetown University’s Center for Security and Emerging Technology, acknowledged that both perspectives on AI in the military hold validity.
“There are numerous potential applications for AI in bureaucratic contexts, which the U.S. military is actively exploring,” Toner said.
Lt. Gen. Michael Conley, commander of Air Force Special Operations Command, informed a congressional committee in May that his units utilized AI “bots” to swiftly convert top secret intelligence to a secret classification, facilitating better communication with drone operators during the Iran conflict.
There’s no denying that AI is also aiding the military in identifying and targeting locations.
A case study released by Toner’s center revealed how the Army’s 18th Airborne Corps employed AI for artillery strikes “as efficiently as the best unit in recent American military history,” requiring 2,000 fewer personnel.
“Human operators continue to make essential decisions, but AI … allows for operations at unprecedented speed and scale,” she remarked.
The debate over AI integration within the military, covering control and ethical considerations, has been quite public throughout the Trump administration.
Hegseth and Anthropic are currently engaged in a contentious contract dispute stemming from the company’s concerns regarding unchecked military use of its technology, particularly regarding fully autonomous drones and AI-assisted surveillance that could monitor dissent.
After CEO Dario Amodei stood firm on his concerns regarding the chatbot Claude’s use in classified Pentagon environments, both Trump and Hegseth accused Anthropic of jeopardizing national security.
The Pentagon subsequently classified the San Francisco company as a supply chain risk — terminating its $200 million defense contract and barring other government contractors from collaborating with them.
Anthropic has filed a lawsuit, alleging that the Pentagon is retaliating illegally by applying a designation intended to safeguard against foreign sabotage of national security systems. In the meantime, the Pentagon has shifted its focus to rival tech firms, including Google, OpenAI, and SpaceX, to acquire AI technology that can “augment warfighter decision-making in complex operational environments.”
Toner, a former OpenAI board member who was ousted after a disagreement with CEO Sam Altman, noted that “the general public often seems to underestimate the caution with which the U.S. military approaches new technologies.”
“Commanders aim for mission success, which necessitates the ability to deliver lethal effects efficiently while also avoiding unintended consequences like friendly fire or civilian casualties,” she concluded.
Also Read: Israeli forces make historic push inside Lebanon and complicate an Iran deal
