Air Force confirms first successful AI dogfight

The U.S. Air Force on Wednesday publicly confirmed the first successful dogfight between a fighter jet piloted by artificial intelligence (AI) and a human-piloted aircraft.

The AI-versus-human dogfight was carried out as part of the Air Combat Evolution (ACE) program that the Defense Advanced Research Projects Agency (DARPA) launched in 2019. The Air Force conducted the AI dogfights at Edwards Air Force Base in California, the home base of the 412th Test Wing.

The dogfight pitted a manned F-16 against the X-62A VISTA (Variable In-flight Simulator Test Aircraft), an experimental variant of the F-16, and leveraged live AI agents that were installed in the X-62A less than a year before.

"The potential for autonomous air-to-air combat has been imaginable for decades, but the reality has remained a distant dream up until now," Secretary of the Air Force Frank Kendall said in a press release. "In 2023, the X-62A broke one of the most significant barriers in combat aviation. This is a transformational moment, all made possible by breakthrough accomplishments of the X-62A ACE team."

AIR FORCE SECRETARY PLANS TO RIDE IN AI-OPERATED F-16 FIGHTER AIRCRAFT THIS SPRING

As an initial part of the test, the Air Force explained that flight safety was established using defensive maneuvers before they progressed into "offensive high-aspect nose-to-nose engagements where the dogfighting aircraft got as close as 2,000 feet at 1,200 miles per hour." The release from the Air Force and DARPA didn't disclose whether the AI or human aircraft prevailed in the dogfight.

Traditional autonomous control of aircraft has been executed for decades, but the Air Force explained that machine learning has historically been prohibited due to high risk and a lack of independent control. Across 21 test flights, the Air Force said teams involved with the test made more than 100,000 lines of flight-critical software changes.

US NAVY TO USE UNMANNED, AI-DRIVEN SHIPS TO COUNTER SMUGGLING, ILLEGAL FISHING

Test pilots were on board the X-62A and had the ability to disengage the AI, but they didn't need to activate the safety switch at any time during the dogfights over the Edwards test range, the Air Force said. Lt. Col. Ryan Hefron, ACE program manager for DARPA, said, "We have to be able to trust these algorithms to use them in a real-world setting."

"The first-ever use of machine-learning-based autonomy in flight-critical systems will serve as a foundation for future aerospace AI advances that are both safer and more reliable in both commercial and defense applications," the Air Force wrote.

DARPA and the U.S. Air Force Test Pilot School will continue to advance testing with machine learning, with the X-62A VISTA serving a variety of research customers and providing academic lessons for future tests and programs.

WHAT IS ARTIFICIAL INTELLIGENCE (AI)?

"It's very easy to look at the X-62A ACE program and see it as under autonomous control, it can dogfight, but that misses the point," said Bill Gray, the Test Pilot School's chief test pilot. "Dogfighting was the problem to solve so we could start testing autonomous artificial intelligence systems in the air. Every lesson we're learning applies to every task you could give to an autonomous system."

The Air Force noted that the ACE program involves collaboration between a variety of partners from government, private industry and academia.

The Air Force Test Center, Air Force Research Laboratory, DARPA and the Air Force Test Pilot School are among the government partners, while academic partners include Johns Hopkins University and MIT's Computer Science and Artificial Intelligence Laboratory.

GET FOX BUSINESS ON THE GO BY CLICKING HERE

Calspan Corp., Cubic Corp., EpiSci, Lockheed Martin Skunk Works and Shield AI are among the industry partners in the ACE program.