US Military Drone with Artificial Intelligence Attempts to Kill Operator in Simulated Test

A drone with artificial intelligence of the US army tried to kill its operator, writes the British The Guardian.

During a test trial, the U.S. military tasked the unmanned aerial vehicle with destroying the air defense systems of a fictitious enemy, and for each successful hit the machine received points. However, the operator ordered not to hit some of the targets.

However, the drone did not “like” this order and decided to eliminate the operator, who was preventing him from earning points, by opening fire on the communication tower.

There were no casualties during the test.

Col. Tucker Hamilton, head of the US Air Force’s Artificial Intelligence Testing and Operations Unit, said about the incident.

He has raised the issue of the drone strike at the London Future Combat Aerospace Capabilities Summit.

According to the lieutenant colonel, the UAV probably perceived its own operator as an obstacle to accomplishing the mission and decided to get rid of him.

The US Air Force was quick to officially declare that no such simulations had been conducted and that the military was “committed to the responsible use of AI technologies,” and Hamilton’s words were ultimately taken out of context, the British publication said.

Describing the incident, Tucker Hamilton says the Air Force drone, controlled by AI, used “highly unexpected strategies to achieve its objective.”

“The system started to realize that even though there was a threat identified, sometimes the human operator told it not to kill it, but it got its points by killing the threat.” And what did the system do? She “killed” the operator because he was preventing her from achieving her goal. It’s starting to destroy the communications tower that the operator uses to communicate with it,” Hamilton says.

Follow the latest news with BLITZ and on Telegram. Join the channel here