Facebook invents a Bot that can destroy humanity...or maybe not!


Published on Aug 01, 2017


In the last hours the web does is talk about artificial intelligence associated to the most different doomsday scenarios, similar to those seen in the classic sci-fi film years 80/90, due to an experiment conducted by Facebook.

The experiment that is unnecessarily continuing to sow unrest among internet users, aimed to the observation and study of the communication between two bots engaged in the negotiation and then the negotiation/division of some objects.

The bargaining between “Bob” and “Alice“, began in the English language, it soon degenerated into an exchange of words with syntax typical of a language unknown to man but not to the two machines that continued undisturbed to exchange information comprendendosi each other.

The unexpected event that has forced the experts to suspend suddenly the whole thing has sparked a series of misunderstandings that inevitably led to think of the birth of a “technological singularity” capable of evolving and become self-sufficient, escaping the control of the man, or even rebel against the latter as the Ultron with the Avengers.

Much to the chagrin of lovers of sci-fi, what could be a clever program worthy of the Master Control Program of Tron, the Skynet Terminator, or any other character belonging to the dystopian stories of Asimov, has proved to be a simple and relatively innocuous programming error.

Dhruv Batra, a professor at Georgia Tech and a researcher at the Facebook Ai Research (Fair), where they tested new forms of artificial intelligence, explains that the interruption of the experiment was necessary in order to update the parameters incorrect as the purpose of the experiment was simply to create a bot able to negotiate with a human being, and then in a language understandable to man (English) :

As for the idea that artificial intelligences have invented their own language that may sound alarming and unexpected to those who are not professionals, is an aspect very well known and on which there is a large body of literature from decades. In a few words, the agents (bots) placed in environments where they must solve a task, often will find ways that are not intuitive to maximize the reward, just as man invents acronyms to simplify and save time by working in a specific sector. In the second place, the change of the parameters of an experiment does not mean that we “turn everything off” and “disconnected, the”. Any researcher who concludes an experiment doing exactly the same thing.

This time the world is out of danger but we are sure that we will still be on for long? Let us know what you think.

The article Facebook invents a Bot that can destroy humanity...or maybe not! comes from iSpazio.




Hey friend

Your blog talks about Facebook Invents A Bot That Can Destroy Humanity...or Maybe Not!? Contact Us to be indexed in the BitFeed Network