Skynet is a technology reached by the Terminator franchise. In this saga, Skynet is an artificial intelligence which thought that the Human race is malicious, and has to be exterminated (pretty optimistic…). This artificial intelligence has been developed by engineers to help Humans for some difficult situations but, obviously (we are in a science-fiction movie), the artificial intelligence has developed itself and became aware of itself. People who are afraid of artificial intelligence always bring up this example, whenever there is an artificial intelligence success. But, is Skynet staying at the state of fiction, or this technology can be realize today?
Artificial intelligence is a way to made his own computer to do something without explicitly programming it. John McCarthy has introduced the term in 1955. He defined the artificial intelligence as “the science and engineering of making intelligent machines”. The artificial intelligence science is vast. It can be splitted into subdomains like mathematical optimizations, multi-agent systems or machine learning. The current AI’s subdomain which reach is machine learning. Arthur Samuel defined in 1959 that machine learning is a “Field of study that gives computers the ability to learn without being explicitly programmed”.
A computer is a deterministic tool. This one will executes instructions you’re giving to it, and it will return to you a message or something else like a graphic in your screen. Machine learning gives the opportunity to give data to a computer, or allows this one to make some things in a based-environment, to maximize his chance to successfully make a regression on data or classify data (for example). To do that, the computer will iterate on data, try to learn characteristics from data, and minimize the number of errors he made by classifying his data. And this is an algorithm!
Stephen Hawking, the famous theoretical physicist and cosmologist, says “The development of full artificial intelligence could spell the end of the human race”. In the same article, Thomas G. Dietterich says that the real threat is “autonomous”. But the most interesting part of his speach is this following sentence: “AI systems will not become spontaneously autonomous, they will need to be designed that way”.
“They will need to be designed that way”
Ass I said, an artificial intelligence canno’t be aware to make some evil things if the algorithm has not been designed for that. A computer listen and execute some instructions, and has to be programmed to do this thing.
We are afraid and proud about artificial intelligence success, but are we far to build a Skynet’s clone? We don’t have to forget that an artificial intelligence is developed to perform one task, like playing go. So, is a programmer, in this world, is currently trying to code a piece of program that will take control of everything? Maybe! But, to be optimistic, I think there is more chance that the Human race will kills itself that an artificial intelligence will takes control of everything…