Artificial intelligence defeated US fighter pilot in a clean sweep of simulated battles
An AI pilot has defeated a US Air Force pilot in a virtual F-16 dogfight in a “coming of age” moment for artificial intelligence.
The US military’s AlphaDogfight Trials was organised by the Defense Advanced Research Projects Agency (Darpa) – a secretive branch of the US Department of Defense responsible for the development of futuristic technologies.
It sought to demonstrate the “feasibility of developing effective, intelligent autonomous agents capable of defeating adversary aircraft in a dogfight.”
The winning AI pilot, developed by Heron Systems, defeated other AI adversaries before going on to beat a human pilot wearing a VR helmet by a score of 5 – 0 in the final.
“We’ve gotten an opportunity to watch AI come of age [against] a very credible adversary in the human pilot,” said Col. Dan Javorsek, program manager in Darpa’s Strategic Technology Office.
“The AlphaDogfight Trials is all about increasing trust in AI. If the champion AI earns the respect of an F-16 pilot, we’ll have come one step closer to achieving effective human-machine teaming in air combat.”
The human pilot, who went by the name ‘Banger’, said that he was unable to match twisting techniques adopted by the AI pilot that he had not witnessed in human-to-human air combat.
“Standard things we do as fighter pilots are not working,” he said during a live broadcast.
AI pilots have a significant advantage over human pilots, as they are not affected by the extreme G forces that occur when manoeuvring at high speeds.
They are also able to aim and fire to a superhuman level, though until now artificial intelligence has lacked the tactical thinking that humans are capable of.
The AI system was developed through deep reinforcement learning in order to overcome this and defeat the human pilot.
Darpa said the AlphaDogfight Trials is a precursor to its ACE program, which ultimately aims to use AI algorithms to fly real aircraft.