One Reason Elon Musk is worried about Artificial Intelligence
Everyone has seen it in the movies, but artificial intelligence doesn’t seem that dangerous. Plus all the movies are just movies after all. It’s all fake and we’re very far from that in real life. However, after seeing Google’s AI demo of a computer program learning how to play video games, or learning what a cat was and being able to draw one simply by watching Youtube videos, AI is already becoming quite dangerous. AI is definitely an exciting new field, where if it can be used to help people manage their work day, such as in the movie “Her” or if it could augment voice recognition and search, it would be very helpful and powerful. But what about the movie “Terminator”? Could that ever happen?
If this AI program is put into a robot or a weapon, it could obviously be very dangerous. Something super human that can learn. And we’re really not that far off either. Perhaps 5, 10 years of development, and a Japanese life-like robot later, and we have a Terminator robot. How could a robot like that with incredible senses be stopped? You would have to attack it with something new each time like in the movie The Incredibles, but with the ability to process data and act in microseconds, there’s no reason something like this can’t dodge bullets or knock projectiles out of the sky before it gets hit. Basically it would be no more dangerous than conventional weaponry that exists today, except that it could have a mind of its own.
But would something super human even need to take the form of a human or a robot? AI could live within computer networks. Since we will all be connected to the network in one way or another through devices that we control, or even the internet of things devices, we are giving A a large playing field that it can do some damage with. Elon Musk even suggested that if the AI pointed all Tesla cars that were connected to the network to simply drive to Rhode Island it would cause plenty of havoc on its own with traffic jams.
Even something completely innocent and seemingly harmless that was asked of A could accidentally turn into a chaotic problem. For example, in the completely made up episode of Rick and Morty (an Elon Musk favorite) "Meeseeks and Destroy", and simple and harmless request from an artificial being capable of effectively granting wishes upon command, end up turning into a chaotic meltdown situation. While this occurs in a fantasy world, the writer certainly draw parallels of how making a request from an artificial intelligent system in real life could easily have adverse and unexpected effects. Humans simply don't completely understand AI and likely never will.