A US Air Force F-16 pilot just battled AI in 5 simulated dogfights, and the machine emerged victorious every time
An AI algorithm piloting an F-16 Fighting Falcon in a simulated dogfight against a seasoned US Air Force pilot achieved a flawless victory with five straight wins in a fierce competition Thursday, according to multiple reports.
The Defense Advanced Research Projects Agency (DARPA) held the final round of its third and final AlphaDogfight competition Thursday, putting an AI system designed by Heron Systems against a human pilot in a "simulated within-visual-range air combat" situation.
Heron's AI went head-to-head with a graduate of the Air Force's Weapons Instructor Course with the callsign "Banger," according to Breaking Defense. A expert commentator, DARPA's Justin Mock, said that the AI algorithm demonstrated "superhuman aiming ability" during the dogfight. During the fight, the human pilot never scored a single hit.
Breaking Defense characterized the simulated engagement as a "one-on-one combat scenario" in which combatants fired "forward guns in a classic, WWII-style dogfight," indicating that while artificial intelligence is rapidly progressing, more work is required before it is ready for the complexities of modern warfare.
Ben Bell, Heron's Senior Machine Learning Engineer, told Sandboxx News that the simulated environment put the AI combatant at an advantage over its human foe. That being said, the virtual environment also allowed the human pilot to execute high-G combat maneuvers that might not be possible for a human pilot in the physical world.
Regardless, the achievement is significant for an AI system for which development, according to Breaking Defense, began less than a year ago. Bell said that Heron's AI program learned fast, gaining roughly "12 years of experiences" over the course of 4 billion simulations, Defense One reported.
Heron's system competed against the human pilot after defeating other AI algorithms in previous rounds of the competition.
Last year, DARPA explained that while AI can beat human beings in games like chess, there is no AI currently in existence that "can outduel a human strapped into a fighter jet in a high-speed, high-G dogfight." The agency is trying to change that and advance AI for future manned-unmanned teaming.
The AlphaDogfight is aimed at moving DARPA's Air Combat Evolution (ACE) program forward.
The ACE program, according to DARPA, is designed to "deliver a capability that enables a pilot to attend to a broader, more global air command mission while their aircraft and teamed unmanned systems are engaged in individual tactics."
Commenting on the ambitious goals of DARPA's ACE program, Col. Dan Javorsek, the program manager in DARPA's Strategic Technology Office, said last year that the agency envisions "a future in which AI handles the split-second maneuvering during within-visual-range dogfights, keeping pilots safer and more effective as they orchestrate large numbers of unmanned systems into a web of overwhelming combat effects."
Given the Department of Defense's growing interest in AI, DARPA is not the only one pursuing these types of opportunities.
Speaking at a Mitchell Institute event in early June, Air Force Lt. Gen. Jack Shanahan, then the head of the Joint Artificial Intelligence Center, said that the Air Force is looking at putting an AI-driven autonomous drone up against a manned fighter aircraft in a dogfight next year.
It is unclear what next year's air-to-air battle will look like, assuming it happens. It is also unclear exactly what the AI-driven autonomous drone might look like.
https://www.businessinsider.com/ai-just-beat-a-human-pilot-in-a-simulated-dogfight-2020-8