Autonomous Fighter Jets Use AI in Real World Tests Against Human Pilots

Military researchers have successfully conducted real-world trials on autonomous fighter jets, pitting AI-powered jets in dogfights against human-piloted F-16s.

The Defense Advanced Research Projects Agency (DARPA), the Pentagon’s emerging tech research arm, has been exploring integrating AI into fighter jets to improve pilots’ capabilities in combat scenarios through its Air Combat Evolution (ACE) program.

In the latest update, DARPA conducted in-air tests, with an AI system autonomously flying a fighter jet in engagements against human-piloted planes.

Existing F-16 fighters were retrofitted for autonomous flight to form the X-62A or VISTA (Variable In-flight Simulator Test Aircraft).

The X-62A jets are powered by a machine learning-based system built on historical flight data to make informed decisions for flight and combat scenarios.

DARPA said the AI-powered jets “show transformational progress for human-machine teaming and trusted autonomy.”

“The X-62A demonstrated that cutting-edge machine learning-based autonomy could be safely used to fly dynamic combat maneuvers,” said Frank Kendel, secretary of the Air Force.

“The team accomplished this while complying with American norms for safe and ethical use of autonomous technology.”

Older fighters required human pilots to have pinpoint accuracy to shoot down enemy planes but modern jets come equipped with an arsenal of tech tools to help in combat situations, including sensors, radar and guided missiles.

Related:NASA Autonomous Flight Software Could Be Flying Vehicle Game Changer

DARPA wants AI to further improve the combat effectiveness of fighter jets. The research agency also wants to ensure human pilots trust the system.

“We have to be able to trust these algorithms to use them in a real-world setting,” said Lt. Col. Ryan Hefron.

DARPA’s latest tests were conducted in the skies above Edwards Air Force Base in California in 2023 and are set to continue this year.

Previously, ACE program tests were conducted in virtual environments with pilots wearing virtual reality headsets to take on AI-powered planes.

DARPA researchers argue, however, that more than virtual environments are needed to fully test AI systems because they cannot replicate the unpredictability and complexity of real-world interactions.

“One of the primary problems machine learning and AI has had to overcome is something called the sim-to-real problem,” said Chris Cotting, director of research at the U.S. Air Force Test Pilot School.

In tandem with ACE, the Air Force is also conducting the Skyborg project, which is attempting to develop unmanned combat aerial vehicle systems.

Related:FAA Approves Autonomous Flight System Testing

There is also the Air Force’s Wingman project, where a swarm of missiles-carrying unmanned drones fly alongside a fighter jet to assist in combat scenarios.

U.S. defense forces aren’t alone in trying to develop autonomous fighter jets.

Chinese military researchers at the Aerodynamics Research and Development Center in Mianyang conducted similar tests last summer and Japan says it wants unmanned fighter jets by 2035.

This article first appeared in IoT World Today's sister site, AI Business.