Facebook Says Its AI Bots Developed Their Own Language

Facebook said the chatbots it developed invented their own artificial language with no human input.

The Atlantic details the findings in a Facebook report, which explains how researchers at the Facebook Artificial Intelligence Research lab discovered how their bots began talking to each other in a new language.

The conversations between bots "led to divergence from human language as the agents developed their own language for negotiating," the report reads.

The episode forced the Facebook developers to change one of the models they used to a fixed supervised model.

The findings are significant because they prove that forms of artificial intelligence have the ability to communicate in their own language with one another.

Much has been written about artificial intelligence in recent years, with some experts warning about the dangers they pose to civilization.

• Scientists Use Artificial Intelligence to Catch 'Criminals' by Reading Minds
• AI Might Be Better at Predicting Heart Attacks Than MDs