The U.S. military has quietly rewritten part of how it decides what to strike in battle, opening the door to a much larger role for artificial intelligence in targeting.
According to Katrina Manson of Bloomberg, the Department of Defense has revised its classified targeting doctrine to envision "systems where AI initiates actions with human monitoring." In plain terms, that describes setups where software could begin the targeting process on its own, with people watching over the loop rather than driving every step.
Reuters-style coverage aggregated across outlets — including reports surfaced by MSN, Crypto Briefing, and Seeking Alpha — frames the change the same way: the Pentagon is opening the door to a far more active AI role in combat decisions. The reporting describes this as another step in the military's broader push to fold AI into how it fights, not a one-off policy tweak.
The key shift is subtle but significant. Past framing of military AI has leaned heavily on keeping a human firmly "in the loop" — making the actual decisions. The newly described language, by contrast, points toward AI taking the first action while humans monitor, a posture often called "human on the loop." That distinction matters because it changes who, or what, moves first.
The doctrine itself is classified, so the public details come from reporting on the document rather than from the full text. The sources do not specify exactly which weapons or scenarios are covered, what safeguards apply, or when the changes take effect.
Why it matters: how the world's most powerful military draws the line between human judgment and machine initiative in life-and-death decisions sets a precedent that allies, adversaries, and the public will all be watching closely.