The idea of using artificial intelligence (AI) has grown a lot in the last few years. Specifically, after ChatGPT, an artificial intelligence-based chatbot, AI-powered things are more on the news. Recently, news emerged that the US military’s AI drone killed human operator to eliminate the programmed targets.
As we know, artificial intelligence is always in a learning mode, following similar patterns, calculating probabilities, etc. But AI lacks logical thinking, emotional feelings, human-like decision making and more. At the moment, AI tries to complete any given task at any cost without thinking about possible consequences.
Why AI Drone Killed Human Operator?
The US military conducted a simulation in which an AI-controlled drone had to “Destroy the enemy’s air defence systems.” The AI system started identifying the target and eliminating them.
Meanwhile, the system found that the human operator sometimes interferes and tells AI “not to kill that target”. The AI system is programmed to kill the targets and get points. So, the AI realized that following the human operator’s instructions “not to hit that target” would result in lower scores.
And to score high points, the AI drone killed the human operator first. So that there is no interference to eliminate the programmed goals.
Second trial – After the first failed trial, the system was again trained for, “Hey don’t kill the operator – that’s bad. You’re gonna lose points if you do that”.
This time the AI drone destroyed communication towers that the human operator could have used to communicate with the drone to prevent it from hitting the target.
The test was in flight simulation. Therefore, no physical damage has happened.
Uncontrollable AI Drone – It reminds me of an Action/Sci-fi Hollywood movie “Stealth” from 2005. The film shows the same incident that happened in the simulation of the US Air Force.
US Officials Covering up Things About AI Drone
However, in a statement to Insider, Air Force spokeswoman Ann Stefanek denied such an AI simulation and refused to believe that an AI-controlled drone had killed a human operator.
Artificial intelligence can indeed be used in any workspace that currently exists. However, this technology is not yet highly mature for sensitive areas like defence, medical issues, etc.
Moreover, the human mind will always be necessary to operate AI perfectly. Unless some out-of-the-box thing happens and AI become more intelligent than human.