Toggle Menu
  1. Home/
  2. Tech & Science/
  3. IT&C/

Facebook’s artificial intelligence accidentally developed its own language

Scientists from Facebook’s artificial intelligence laboratory made an unexpected discovery while trying to improve the activity of the chat bots. The bots, called “dialogue agents”, created their own language.

By using a series of machine learning algorithms, dialogue agents were left to talk freely, in the attempt of improving their conversational skills.

In the meantime, though, the bots started deviating from scripts and started communicating in a new language, without any human involvement.


Their conversation might not have any sense, but what is interesting is that when given the opportunity, artificial intelligence seems to deviate and create something completely new.

In fact, this isn’t the first time this happened. Scientists from Open AI, one of Elon Musk’s initiatives, forced their own algorithms to communicate between them without telling them how, and each of them understood what needs to be done.

This wasn’t the only discovery of scientists from Facebook. They also saw that the bots can become extremely good negotiators. After learning how to negotiate, the bots used machine learning and advanced strategies in the attempt of improving the final result of the negotiation.

Meanwhile, they became really good at it, and even started simulating the “interest” for something, only to “sacrifice” that thing in a false compromise in advanced negotiation stages.

Daniel Higgson