By Aaron Kesel
If that title didn’t capture your attention, I don’t know what will. No that isn’t clickbait, the military has been working on developing artificial intelligence for some time. But now the machines have surpassed humans, again winning a dogfight – a term used to describe an air battle between jets. And if that’s not enough, the military wants self-healing robots straight out of The Terminator, which by the way was the villain.
According to Defense One, artificial intelligence (A.I.) beat a fighter jet pilot for the second time. The military held a simulation between A.I. pilots and a human pilot, and found that one specific A.I. beat others and the human in five rounds.
The record was broken at the finale of the U.S. military’s AlphaDogfight challenge, which was set up as an effort to “demonstrate the feasibility of developing effective, intelligent autonomous agents capable of defeating adversary aircraft in a dogfight.”
This isn’t the first time that a human fighter pilot was defeated by A.I. in a contest. A 2016 demonstration showed that an A.I. nicknamed Alpha could win against an experienced human combat pilot. However, the recent DARPA simulation pitted a bunch of different AI against one another and then against a human in the end.
If that’s not enough, in the same week as The Mind Unleashed reported, the military suggested creating Terminator-like robots that can heal themselves on their own utilizing a new shape-shifting material.
In fact, Military.com reports that the film’s villain, the T-1000, even directly provided the inspiration to one of the Army engineers working on the project to develop “soft robotic” drones and unmanned aircraft based on flexible, self-repairing and self-reconfiguring materials.
“We want a system of materials to simultaneously provide structure, sensing and response,” said Frank Gardea, an aerospace engineer at the Army Combat Capabilities Development Command’s (CCDC) Army Research Laboratory.
The project is being worked on at the U.S. Army’s Aberdeen Proving Ground facility in Maryland. So far they have developed a 3D-printed, flexible polymer with a “unique shape memory behavior” that can be programmed to remember which will snap back to certain shapes,