A sharp debate is unfolding inside Washington over who controls America's increasingly autonomous weapons—and where the line between human judgment and machine decision-making should be drawn.
The U.S. Senate is pushing the Department of Defense to stand up an entirely new military command dedicated to unmanned and robotic systems. According to DefenseScoop, the proposed Robotic and Autonomous Systems Combatant Command would carry special authorities for testing, evaluation, and streamlined acquisition—essentially a fast lane for getting autonomous capabilities into the field.
At the same time, a senior military voice is pumping the brakes on how far that autonomy should go. Admiral Frank Bradley, the U.S. Special Operations Commander, has warned that artificial intelligence should not be permitted to make lethal decisions on the battlefield. Humans must retain control over the use of force, he argued, drawing a firm line between AI as a tool and AI as a decision-maker.
Meanwhile, analysts are urging policymakers to rethink how they define autonomous weapons in the first place. Writing for CSIS, Kateryna Bondar argues that the Pentagon's current policy revision must look beyond the physical platform—the drone or missile—and encompass the software orchestration layer that actually directs it. In her view, software, not hardware, will determine the character of the next war.
Together, these developments reveal a system under strain: the military wants to move faster on autonomous capabilities, Congress wants a new institutional structure to manage them, and top commanders are warning that speed cannot come at the cost of human accountability.
How the U.S. resolves this tension—between rapid deployment and meaningful oversight—will set a precedent that allies and adversaries alike are watching closely.