USAF Tests AI in Warplanes

Can AI effectively fly combat aircraft? The United States Air Force is testing that thesis with the XQ-58, an experimental airplane codenamed Valkyrie.

The XQ-58 is a joint collaboration between the Air Force Research Lab and defense contractor Kratos to determine whether AI-powered drones, dubbed collaborative combat aircraft, can be effective in combat.

While the XQ-58 doesn’t have a pilot, it would be fitted with missiles and bombs, as well as data-gathering sensors. The plane could be used for certain high-risk tasks, such as taking out air defense sites deep in enemy territory.

The idea is to surround a manned aircraft, such as the F-35 or F-22, with a swarm of collaborative combat aircraft. The drones would not fire their weapons on their own, but would be under the control of a pilot nearby.

The core issues is the escalating costs of fighter jets. The F-35 Lightning II, the Air Force’s go-to fighter, costs $80 million apiece, while the F-22 Raptor, the force’s air supremacy jet, costs $143 million each. That has led to the Air Force scaling back the size of its fighter jet fleet to the lowest it’s ever been.

Drone aircraft would be much cheaper to build, costing anywhere from $3 million each for disposable aircraft, up to $25 million for more sophisticated planes, such as XQ-58. This would give the Air Force “affordable mass,” according to an article this week in The New York Times.

New York Times reporter Eric Lipton was invited to Eglin Air Force Base to observe the XQ-58 program. Following a test flight in which F-15 pilot Major Ross Eldor flew alongside the Valkyrie, the Major expressed some researvations about the program. “It’s a very strange feeling,” Major Elder told the New York Times. “I’m flying off the wing of something that’s making its own decisions. And it’s not a human brain.”

Whether AI can replicate the actions of an experienced pilot has yet to be seen. Dubbed “tactical autonomy,” the Air Force Research Laboratory has invested millions of dollars to test airframes and algorithms for pilotless flight, including the Valkyrie, the MQ-20 Avenger, the UTAP-22 MAKO, and the VISTA X-62.

The Valkyrie is the latest collaborative combat aircraft to be tested. Last month, the AFRL announced a successful three-hour training flight of the Valkyrie and its AI and machine learning powered systems.

“The mission proved out a multi-layer safety framework on an AI/ML-flown uncrewed aircraft and demonstrated an AI/ML agent solving a tactically relevant ‘challenge problem’ during airborne operations,” Col. Tucker Hamilton, DAF AI Test and Operations chief, said in a press release. “This sortie officially enables the ability to develop AI/ML agents that will execute modern air-to-air and air-to-surface skills that are immediately transferrable to other autonomy programs.”