The world’s first known fight between a human pilot and an AI-controlled fighter jet took place in California, the US military announced.
During an exercise over Edwards Air Force Base, the pair of F-16 fighter jets flew at speeds of up to 1,200 mph and came as close as 600 yards during A dogfight, also known as aerial combat.
One was manned, while the other was a modified version of the F-16, called the X-62A, or VISTA (Variable Flight Simulator Test Aircraft).
In flight, the AI The algorithm relies on analysis of historical data to make decisions about present and future situations, according to the Defense Advanced Research Projects Agency (DARPA), which conducted the test.
This process is called “machine learning” and has been tested for years in ground simulators, said DARPA, a research and development agency of the US Department of Defense.
In 2020, so-called “AI agents” defeated human pilots in simulations in all five of their showdowns – but the technology was to actually be used in the air.
The pilots were on board the X-62A in case of emergency, but did not need to return to the controls at any time during the test dogfight, which took place in September last year was announced this week.
The result represents a “transformational moment in aerospace history,” DARPA said in a statement.
He did not reveal which plane won the dogfight.
“The potential for autonomous air-to-air combat has been imaginable for decades, but the reality has until now remained a distant dream,” said Secretary of the Air Force Frank Kendall.
“In 2023, the X-62A broke one of the most important barriers in combat aviation. This is a transformative moment, made possible by the groundbreaking achievements of the X-62A ACE team .”
Learn more:
Are we heading towards World War III?
The new technology that brings loved ones back to life thanks to AI
Colonel James Valpiani, commander of the US Air Force Test Pilot School. said: “Dog fighting is a perfect case for the application: machine learning.
“Aerial combat is extremely dangerous. So if machine learning can work effectively in an environment as dangerous as aerial combat, it has great potential to gain the trust of humans as we move toward less dangerous but less dangerous applications. equally complex.
He added: “The X-62A is an incredible platform, not only for research and test advancement, but also for preparing the next generation of test managers.
“By ensuring that the capability it has is safe, efficient, effective and responsible, the industry can view the results of what the X-62A ACE team has done as a paradigm shift.
“We fundamentally changed the conversation by showing that this can be done safely and responsibly.”