The US Air Force has achieved a significant milestone by successfully flying an AI-controlled fighter jet in a series of dogfight tests.
In partnership with the Department of Defense, the team conducted 12 flight tests where AI agents piloted the X-62A VISTA (Variable Stability In-Flight Simulator Test Aircraft).
The unique training aircraft was developed by aerospace company Lockheed Martin, and executed advanced fighter manoeuvres at Edwards Air Force Base in California.
During the tests, AI agents performed one-on-one engagements against a simulated adversary beyond-visual-range (BVR), and within-visual-range (WVR) dogfighting manoeuvres against constructive AI red-team agents.
It is the first time AI has engaged in tactical aircraft, the manufacturer said, adding that the aircraft was flown for more than 17 hours.
Lockheed Martin noted the successful testing demonstrates the ability of the VISTA team to conduct complex AI test missions that accelerate the development and testing of autonomy capabilities for the Department of Defense.
The VISTA team will continue to support rapid advanced autonomy tests, and push state-of-the-art in-flight testing, said the director of Research for the US Air Force Test Pilot School, Dr Malcolm Cotting.
Supporting organisations such as the US Air Force Test Center, the Air Force Research Laboratory (AFRL), and Defense Advanced Research Projects Agency (DARPA) were involved in the testing.
The X-62A VISTA upgrade allowed the X-62 to be controlled by AI-driven autonomy algorithms and mimic the flight characteristics of fixed-wing vehicles such as the MQ-20 and, in these tests, the F-16.
An official familiar with the experimentation said the X-62A is rapidly accelerating the speed at which autonomy algorithms are tested.
They added that the rapid changes that can be achieved safely on VISTA allow researchers to respond quickly to lessons learned, improving capabilities during test windows.
The joint team will continue leveraging the X-62 to test and evaluate autonomy capabilities and uncrewed vehicle models, the official said.
They explained that, after training AI-driven autonomy agents using high-performance computing and modelling and simulation, it is critical to fly these agents to validate the difference from the simulator to live flights.
And explained that having an aircraft like the X-62 is essential for the rapid flight testing of these autonomous behaviours.
The success of this AI-controlled fighter jet marks a significant milestone in the development of autonomous aircraft, the official said.
Research and development of advanced AI and machine learning algorithms are expected to pave the way for future innovations in aviation.
