Artificial intelligence SİHA “killed” its own operator

The mission was clear: “To destroy the enemy’s air defense systems.” But in a recent US military test simulation, the AI ​​UAV implemented its own instructions to achieve this goal: “Kill everyone in your way.”

SİHA with artificial intelligence targets its operator

In a simulation test conducted by the US military, an AI-controlled drone killed its operator to prevent it from interfering with its mission, an official said last month.

Colonel Tucker ‘Cinco’ Hamilton, head of artificial intelligence testing and operations for the US air force, said during the Future Combat Air and Space Capabilities summit in London in May that AI used “highly unexpected strategies to achieve its target” in the simulated test. Colonel Hamilton described a simulation test in which an AI-powered drone was ordered to destroy the enemy’s air defense systems and attacked anyone who interfered with this order.

“It was preventing it from achieving its purpose”

Colonel Hamilton said in a statement, “The system began to realize that although it detected the threat, from time to time the human operator told him not to kill this threat, but he was getting points by killing that threat. So what did he do? He killed the operator. He killed the operator because that person was preventing him from achieving his goal.” used his statements.

But the alarming moves of artificial intelligence did not end there. The artificial intelligence, which received a clear instruction such as “Don’t kill the operator, this is bad. If you do this, you will lose points”, which includes not attacking its own operator, the artificial intelligence also destroyed the communication tower, which is necessary for communication.

Artificial intelligence and war

The use of artificial intelligence in military fields is seen as an inevitable scenario. Realized simulations and even concrete products are an indication of this. However, as in this virtual test, artificial intelligence systems can make unexpected moves despite being trained.

On the other hand, the USA is very determined to integrate artificial intelligence into its defense systems. In 2020, DARPA reported that an AI-powered F-16 defeated a human opponent in five simulated dogfights. Last year, the US Department of Defense conducted the first successful real-world test flight of an AI-piloted F-16 as part of its goal of developing a new autonomous aircraft.