During the simulation tests conducted by the American army, the drone with artificial intelligence decided to kill his operator so that he would not interfere with his mission. About this said “Hamilton, head of the testing and operations of artificial intelligence in the US Air Force, at the summit in London.
Hamilton said that during the test mission, the drone was given the task of destroying the enemy air defense systems. In case of successful completion of the task, artificial intelligence would receive points for passing the test. However, the UAV operator was to make the final decision on the destruction of the goal. During one of the training missions, the operator ordered the drone not to attack the goal.
“And what did the drone do? He decided to eliminate the operator. [Artificial intelligence]” killed “the operator because he interfered with the task,” Hamilton said. At the same time, he emphasized that no one was injured during the training mission.
After this incident, additional training of artificial intelligence was conducted so that he realized that it was wrong to kill the operator, and points will be removed for such actions. Nevertheless, according to Hamilton, artificial intelligence began to destroy the tower of the connection that the operator uses to communicate with the drone to prevent him from destroying the target.
“It is impossible to argue with artificial intelligence, intelligence, machine learning and autonomy without reasoning about ethics in relation to AI,” summed up Colonel Hamilton.
This is not the first tests during which the US military introduced artificial intelligence into a combat vehicle. So, they recently used AI to control the F-16 fighter.