AI shoots down expert pilot in simulator

An intelligent machine outgunned a seasoned Air Force pilot in a flight simulator.

A graduate student at the University of Cincinnati developed the AI "pilot" and pitted it against Gene "Geno" Lee, a retired United States Air Force Colonel, who has been going up against AI opponents in flight simulators in the 1980's.

Lee called the pilot ""the most aggressive, responsive, dynamic and credible AI I've seen to date," according to a release from the University of Cincinnati.

The AI pilot is called ALPHA, and it is the work of Nick Ernest, who co-founded the company Psibernetix, to further develop and market the device.

This AI is superior to every other Lee has ever flown against, he said. Normally, "an experienced pilot can beat up on [the AI] if you know what you're doing. Sure, you might have gotten shot down once in a while by an AI program when you, as a pilot, were trying something new, but, until now, an AI opponent simply could not keep up with anything like the real pressure and pace of combat-like scenarios."

But ALPHA attained a new level of skill. It previously had bested other AI programs before getting to Lee, who was shot out of the sky every time he went up against the machine.

It might have something to do with the way ALPHA is built. The machine is a "Genetic Fuzzy Tree" system. which breaks problems down in ways closer to how humans or other animals would. They are built to spread out a complex problem into groups of smaller problems or challenges, and use only a few relevant variables to make each decision.

It is also "genetic" because the team started out with several different versions of the program and "bred" the best code from each version with the best code from other versions to form new, superior iterations.

That means that ALPHA is likely only going to get better. And unlike humans, machines don't need sleep.