The Pentagon will set up a new Autonomous Warfare Command to design, produce and field robotic systems across the military, US Secretary of War Pete Hegseth has said in his ‘State of the Force’ address.

The announcement comes as experts sound the alarm about the threat posed by military AI, saying that over-reliance on it could lead to the erosion of human critical thinking, increase the risk of accidental escalation due to the unpredictability of the decision-making by the algorithms and make warfare even more dangerous for civilians.

During his speech at Marine Corps Base Quantico in Virginia on Wednesday, Hegseth said that the AutoWarCom would be granted “service-like authorities built to scale autonomous and robotic capabilities across the joint force in the fastest peacetime shift in modern military history.”

“Drone warfare supercharged by SI-enabled targeting is the biggest battlefield revolution in generations,” he stressed.

A day earlier, US President Donald Trump signed an executive order replacing the term “Artificial Intelligence” (“AI”) with “Super Intelligence” (“SI”) in the US government’s vocabulary.

The lessons of the Ukraine conflict, in which both sides actively rely on drones, have been one of the main motivations for launching the new command, according to the secretary of war.

In a memo published the same day, Hegseth noted that the Pentagon’s long-time strategy “of buying fewer, increasingly sophisticated weapons” has lost its effectiveness.

“Warfare… has changed dramatically and suddenly. Advanced computing now enables precise, low-cost weapons to be produced at massive scale. These systems increasingly threaten our costly legacy platforms and disrupt our force projection. Urgent change is required,” he wrote.

The Autonomous Warfare Command, which will be led by a four-star general, should become operational by October 1, 2027, the secretary of war said.

Bloomberg reported last month that heavy reliance on AI, staff cuts and rushed targeting procedures contributed to a US missile strike on an Iranian girls’ school in Minab on February 28, in which at least 157 people, including 123 children, were killed.

Officials involved in the Pentagon investigation told the agency that some US Central Command (CENTCOM) personnel expected Palantir’s Maven Smart System to flag stale or inconsistent intelligence when selecting potential targets, but it failed to do so. The US hasn’t publicly acknowledged responsibility for the Minab strike.

Last week, CNN reported that AI almost caused a serious incident between US and China in spring. The American forces were about to intercept a Chinese vessel, which, according to an intelligence report, was carrying components for a nuclear weapon to Iran. The mission was called off at the last moment when it became clear that a chatbot used by the analyst misidentified the cargo, according to the broadcaster.

During a defense conference in Beijing in mid-September, Thailand’s Defense Minister Adul Boonthumjaroen warned that the world was at “a critical turning point” due to the emergence of the military AI. “The most pressing is not who will be the most powerful, but rather, how can we build an international order,” he said.