Facebook AI Shut Down Proves Elon Musk Right And Mark Zuckerberg Wrong

In a recent kerfuffle, Tesla CEO Elon Musk shot back at Facebook founder Mark Zuckerberg after the latter called him “irresponsible” for spouting doomsday rhetoric regarding artificial intelligence. In response, Musk said that the programmer’s knowledge of the matter. In a rather ironic twist, Facebook had to shut down an AI project that was threatening to go out of control, which proved Musk’s concerns valid.

Despite what a lot of rabid AI supporters might think, Musk is not against the use of intelligent machines. His cars and space agency are making use of them to great effect. He is simply advocating for better oversight in order to prevent losing control of a potentially destructive tool. When Facebook noticed that its AI bots started speaking in different languages, this is exactly the kind of scenario that Musk was worried about.

It seems the people working on the project had similar concerns because it shut the project down after noticing the shift, Digital Journal recently reported. Supposedly, the chat bots that Facebook’s engineers were pitting against each other to simulate hard negotiation tactics started using a form of English that looked like gibberish.

In the transcript, the bots started to send phrases like "I can i i everything else," and "balls have zero to me to me to me" to each other. According to the reports, these are not nonsense either. The chat bots were simply shortening sentences in order to speed up the process since using full English would be too slow.

Aside from proving that the Tesla boss has a point, this development is also interesting in terms of how machine learning can be conducted in the future. As The International Business Times notes, allowing AIs to simply create a language and system of their own could lead to considerable jumps in speeds regarding the learning process.