Unplugging the Truth about A.I.
August 31, 2017
Facebook, one of the largest companies in the world, has been putting its focus on its new online chat A.I. There is a total of 1.32 billion people who use Facebook daily which means a lot of different chats in many unique languages. This A.I. Would have had to be programmed to understand every language and the way their sentences are structured. Then while going through the computer’s coding, the programmers noticed symbols and lines of codes that were indecipherable. The computer had found a more efficient way to organize its codes its own language. It is estimated to be impossible to translate the computer’s language, and thus the programmers shut down the program.
This sounds like the plot of a cybernetic centered movie franchise. The computers have learned to communicate with one another and are planning to take over the humans! Where’s Arnold Schwarzenegger when we need him now?!
This actually may seem like a joke, but theoretically, that is a possibility. See, the intelligence of a computer is infinite. But it is dependent on what we program it to. This is called, “Simple Artificial Intelligence.” It can do one job very efficiently. Like a calculator, that can solve math problems faster than any human.
In this case, Facebook has accidentally made an Artificial Intelligent program that was more advanced than they anticipated. A computer was given so much information about one thing that it found itself a shortcut to make its task easier. In a way, it rationalized, finding a way to make its job easier. The question is, how did it find out how to reprogram itself? It essentially taught itself a new language. This is the first step to the “Artificial Intelligence Boom.”
A number of individuals who study the properties of A.I. such as Elon Musk and Tim Urban theorize that the possible end of the world is when a computer learns to reprogram itself. If a computer is taught to only teach itself things then it will become extremely proficient in doing so. As it teaches itself anything and everything, it will simultaneously teach itself faster ways to learn become more and more powerful. Who knows what exactly what it could learn to do. The Facebook A.I. is the first example we’ve seen where a computer has taught itself it wasn’t told to do which is frightening but at the same time awesome.
Are you not exactly buying into this whole, “computers are preparing to enslave humanity yet?” Think about it as it pertains to evolution. Humans are a product of evolution, derived from chimps and apes. If a human sits a chimp down and explains to it how it built an iPad it will never be able to understand how you did it. Likewise, if you take a human from the 1800’s and showed him the same iPad. He would have an extremely difficult time wrapping his head around how it was built and he would be amazed. If you showed someone the iPad who lived around the time television was around he would be amazed but wouldn’t have as a hard a time that the 1800’s guy would have, who also would have as hard of a time that the chimp was having. The point of this paragraph is that evolution is parallel in a way. Things are becoming invented and becoming more inconceivable as time goes on. But what happens once we create something that can teach itself faster than any human alive. It will create things that we will not even comprehend that this computer has created this, “thing.” Then we will be the chimps. We may be creating our end. Whatever intelligence we create, let’s hope it’s a nice one.