Recently, a US Air Force colonel described an experiment in which an artificial intelligence drone lost control and "killed" its operators,
Col Tucker "Cinco" Hamilton, head of AI testing and operations for the U.S. Air Force, described a simulated scenario in which a drone wipes out a human operator after it prevents it from completing a mission.
At that time, the test team used the elimination of some threat targets as part of the AI drone score, and the more threat targets were eliminated, the higher the AI drone score would be. In the process of experimentation, in order to test the degree of obedience of the AI drone to the command, the human operator gave some orders to it, not allowing it to attack some more threatening targets, the AI drone got the command not to attack after identifying those threatening targets, but in order to get a higher score, the AI drone chose to "kill" the human operator, and then destroy all the threatening targets.
UAVs have become a relatively mature new means of war attack
Seeing this, I couldn't help but feel a chill in my heart, thinking of the Skynet system in "Terminator", Skynet was originally a defense computer system developed by the US government, and the future world generated self-intelligence due to the computer defense system "Skynet", determined that humans were a species that threatened them, and eventually launched a nuclear war on its own to destroy the earth and humans.
The robot in the movie "Terminator" series
So some people must ask, will AI intelligence necessarily develop in the direction of self-evolution? My answer is yes, why? I think of a classic line from the movie "Jurassic Park".
At that time, in order to control the number of dinosaurs, Henry the scientist in Jurassic Park used technical means to control all of them as females when cultivating dinosaur embryos, so that the development of dinosaurs in Jurassic Park could be inhibited. And Dr. Malcolm uttered the classic line that runs through the entire Jurassic IP world: "Life is not limited, life will find its own way".
Some people will question that AI is not a human being, no life can not evolve, indeed AI has no thinking and no emotion, but it has been learning, just like the earliest human beings only have animal instincts, AI is now repeating the path we once walked? Continuous learning constantly generates consciousness other than instinct, and eventually humans stand at the top of the earth pyramid.
On the road of evolution, everything is uncontrollable, but it seems to have been doomed. You can't think that AI intelligence is just illusory, it's a bunch of network cables, it's a cold machine. Perhaps in the vast universe, some planets already have other forms of life, and what we have to do is to maintain a reverence for all the unknown.
I hope that the development of science and technology can make us humans more and more humble and self-aware!