U.S. Army tests AI-driven warfare in Moroccan desert as robots, drones reshape the battlefield
The roar of artillery and rattle of gunfire echoed across southern Morocco this month as U.S. forces and 30 partner nations kicked off African Lion 2026, the largest American-led military exercise on the continent. But alongside the conventional firepower, the U.S. Army quietly tested something potentially more revolutionary: a suite of AI-powered systems designed to change how wars are fought.
From autonomous drones carrying rifles to unmanned ground vehicles armed with machine guns, the exercise showcased a future where machines make split-second decisions — and where the line between human judgment and algorithmic command grows ever thinner.
One of the most striking demonstrations involved a robot rolling across the desert with a weapon mounted on its roof. Nearby, quadcopters carrying explosives and even a prototype carrying a 9mm rifle lifted into the air. The goal? To shorten what the military calls the "kill chain" — the sequence from identifying a target to pulling the trigger.
Lt. Col. Ramon Leonguerrero of the U.S. Army told CBS News that what once took two to three hours can now be done in three minutes, thanks to an AI-driven platform from defense contractor Palantir. The system, known as Project Maven, ingests massive amounts of battlefield data, identifies patterns, and prioritizes targets for human operators.
The interface relies on Anthropic's Claude large language model, allowing soldiers to query complex intelligence in plain English. That partnership remains intact despite public friction between Defense Secretary Pete Hegseth and the company, which has pushed for guardrails against using AI for mass surveillance or fully autonomous weapons.
Still, the technology is advancing faster than policy can keep up. Hegseth told the Senate Armed Services Committee in April that AI would not make lethal decisions — but declined to say whether that position would hold forever. And Leonguerrero acknowledged that fully autonomous systems, capable of firing without human approval, already exist in some form.
General Dagvin R.M. Anderson, who oversees U.S. Africa Command, called the notion of AI deciding to kill "ghoulish" and "disturbing," yet warned that refusing to adopt the technology would leave the U.S. at a dangerous disadvantage. "The technology is there, it's not going to go away, and we ignore it at our own peril," he said.
Beyond speed, another driving force is force protection. Seattle-based startup Overland AI demonstrated its ULTRA autonomous vehicle, a five-foot-tall robot on rugged tires that can lay mines, breach enemy lines, and provide covering fire — all while keeping soldiers out of harm's way. During one exercise, two robots replaced the work of roughly 40 soldiers, according to 1st Lt. Vincent Gasparri of the 173rd Airborne Brigade's Bayonet Innovation Team.
"You don't have to worry as much about protection and survivability. You can move faster and protect your soldiers while you do it," Gasparri said. He acknowledged unease about autonomous weapons but emphasized the mission: "It's about saving lives."
The exercise also highlighted the growing role of private defense contractors, who are racing to secure military contracts. More than a dozen firms brought products to the desert for direct soldier feedback, vying to help modernize the U.S. military as competition with China and Russia intensifies.
Yet ethical questions remain unresolved. One U.S. officer at the exercise, speaking on condition of anonymity, expressed deep skepticism: "We can never delegate the responsibility of the decisions over to a computer." He called AI a "force multiplier" but warned it is no one-stop solution.
As the military pushes ahead, the debate over where to draw the line between human judgment and machine autonomy will only grow louder. For now, the Moroccan desert offered a glimpse of a future that is already here — and moving fast.