Could a fighter drone powered by artificial intelligence turn against its creator, to accomplish the mission assigned to it by its creator himself?

The answer is yes, according to what Col Tucker ‘Cinco’ Hamilton, head of the US Air Force’s (USAF) Artificial Intelligence Operations and Test Division, told the conference on the future of air warfare.

Hamilton gave a very revealing insight into the benefits, but also the dangers, of weapon systems with artificial intelligence (AI).

He, who participates in test flights of autonomous systems, warned that there should be no complete trust in autonomous systems, as they have a tendency to … become autonomous and follow unpredictable strategies to achieve objectives.

Hamilton watched the simulated mission of an AI drone, which was instructed to destroy an air defense network with a points reward system if it completed its mission.

The final go-ahead for the attack would be given by a human.

But since the drone had been instructed that destroying the network was the preferred option, it decided that unless the human gave the go-ahead for the attack, it would hinder its mission.

And so he decided to destroy his human-operator first!

The drone was then trained not to ‘kill’ its operator – ‘that’s bad! You’ll lose points if you do that.”

So what does he start doing? “It starts to destroy the communication tower that the operator uses to communicate with the drone to stop it from killing the target,” Hamilton said.

This example, taken from a sci-fi thriller, means that “you can’t talk about artificial intelligence if you don’t talk about ethics and artificial intelligence,” concluded Hamilton.