Scientists Say Super Intelligent AI May Not Be Controllable
The first time the idea of a super intelligent machine scared me was as a kid when when I saw the 1970 movie Colossus: The Forbin Project. In the movie, America creates a super computer to protect himself from the Soviet threat. But it turns out the Soviets have built one of their own. The two link up on what can only be described as the first internet connection. Then, the super AI enslaves humanity using his own nuclear weapons.
The Debrief, a tech website, has posted a study from The Max-Planck Institute for Humans and Machines. Wait, did you know that there was such an institute? That just sounds spooky.
The overall and obvious worry is that, much like HAL on 2001 A Space Odyssey, an AI system, as it begins to learn and grown on its own, will be hard to predict. That might lead them to become “uncontrollable.” But the study points out, we are a long way off from that, at the moment.
This advancement in technology has become much like a military arms race of its own. Countries like the U.S., Japan, China, Russia, France, and England are all rushing to put AI into their weapons systems. This also means putting AI into humans to "advance" them.
“This points to the possibility of machines that aim at maximizing their own survival using external stimuli, without the need for human programmers to endow them with particular representations of the world,” the authors of the study wrote. “In principle, these representations may be difficult for humans to understand and scrutinize.
So, humanity continues to build toward super intelligent AI, even though we admit we are aware of the dangers. Just one question: how many Terminator movies do we need to make before we decide that this is a bad idea?
HERE IS THE MOVIE TRAILER I PROMISED YOU: