Air Force uses AI on military flight for first time

Artificial intelligence controlled sensor and navigation systems aboard a U-2 Dragon Lady reconnaissance aircraft on Tuesday in what the Air Force is hailing as the first military flight with an AI co-pilot.

“Putting AI safely in command of a U.S. military system for the first time ushers in a new age of human-machine teaming and algorithmic competition,” Will Roper, assistant secretary of the Air Force for acquisition, technology and logistics, said in a statement Wednesday. “Failing to realize AI’s full potential will mean ceding decision advantage to our adversaries.”

The U-2 was flown by a human pilot identified by the Air Force with only their call sign, “Vudu,” a major assigned to the 9th Reconnaissance Wing at Beale Air Force Base in California.

The AI algorithm, dubbed ARTUµ in a reference to “Star Wars” droid R2-D2, was responsible for sensor employment and tactical navigation, according to an Air Force news release.

“Though lacking those lively beeps and squeaks, ARTUµ surpassed its motion picture namesake in one distinctive feature: it was the mission commander, the final decision authority on the human-machine team,” Roper wrote in an op-ed for Popular Mechanics published Wednesday.

The pilot and ARTUµ flew a reconnaissance mission during a simulated missile strike. The AI’s main job was to find enemy launchers while the human looked for threatening aircraft, both using the plane’s radar, according to the Air Force release.

After takeoff, sensor control was handed over to ARTUµ, which used lessons previously learned from more than half a million computer-simulated training runs to manipulate the sensor.

During the flight, ARTUµ was pitted against another dynamic computer algorithm in order to prove the new technology, according to the release.

ARTUµ was developed by researchers at Air Combat Command’s U-2 Federal Laboratory and trained to execute specific in-flight tasks that otherwise would be done by the pilot, according to the release. It is a modified version of μZero, a gaming algorithm that bests humans at games like chess and Go, Roper said. 

“Like any pilot, Artuμ (even the real R2-D2) has strengths and weaknesses,” Roper tweeted. “Understanding them to prep both humans and AI for a new era of algorithmic warfare is our next imperative step. We either become sci-fi or become history.”

(2 of 2) Like any pilot, Artuμ (even the real R2-D2) has strengths and weaknesses. Understanding them to prep both humans and AI for a new era of algorithmic warfare is our next imperative step. We either become sci-fi or become history.https://t.co/xQBpyQoqxS @washingtonpost

— Will Roper (@WILLROP3R) December 16, 2020

While ARTUµ was developed for the U-2, it was designed to be “easily transferable” other systems, according to the new release. 

“We know that in order to fight and win in a future conflict with a peer adversary, we must have a decisive digital advantage,” Air Force Chief of Staff Gen. Charles “C.Q.” Brown said in a statement. “AI will play a critical role in achieving that edge, so I’m incredibly proud of what the team accomplished. We must accelerate change and that only happens when our airmen push the limits of what we thought was possible.”