The American agency DARPA (Defense Advanced Research Projects Agency) announced that a combat aircraft controlled by an AI system successfully faced off against a human pilot during a test combat in the air conducted last year.
DARPA began experimenting with the application of AI technology in December 2022 as part of the ACE (Air Combat Evolution) program and developed an AI system capable of autonomously controlling a combat aircraft while adhering to the safety protocols of the United States Air Force.
After conducting combat simulations using the AI pilot, DARPA put it to the test by installing this system in the experimental aircraft X-62A, enabling the aircraft with the AI system to take off from Edwards Air Force Base in California. It was there that the first successful test combat against a human pilot took place, occurring in September 2023.
Human pilots were in the X-62A aircraft with controls that allowed them to override the AI system, but DARPA claims that this was never necessary. The X-62A controlled by the AI pilot engaged in combat against an F-16 fighter flown by a human pilot, and both aircraft demonstrated a “head-to-head” battle, closing in on each other at a distance of 610 meters at a speed exceeding 1,930 km/h. However, DARPA did not disclose the winner of the simulated combat, whether it was the AI pilot or the human.
“The combat was a problem we had to solve in order to begin testing autonomous artificial intelligence systems in the air. Every lesson we learn applies to every task you could give an autonomous system,” said Bill Gray, chief test pilot at the United States Air Force Test Pilot School.
The agency has conducted a total of 21 test flights so far and has stated that testing will continue this year as well. The rapid advancement of AI technology raises concerns about how the military might utilize these systems. The Wall Street Journal reported last year that the Pentagon is attempting to develop AI systems for defense and to enhance its drone fleet.
