Facebook is investing in artificial intelligence research to a potentially creepy end: The social network's top minds want to build software that can do anything from read your status updates to warn you when you're about to upload an embarrassing picture. Put another way, Facebook wants to stop you from uploading drunk selfies.

Let me just say that some of us like uploading drunk selfies. But a lot of users probably won't like the slightly dystopian future when Facebook's robots try to censor their content. The particular branch of AI research that would enable these kinds of features is known as "deep learning." It's an exciting field! However, Facebook's ideas for how it would use deep learning technology just sound a little off.

"Imagine that you had an intelligent digital assistant which would mediate your interaction with your friends and also with content on Facebook," Yann LeCun, a pioneer of deep learning research and the head of Facebook's Artificial Intelligence Research lab, told Wired recently. LeCun also suggested that Facebook might apply deep learning technology to Oculus Rift. It kind of sounds like he wants to apply deep learning technology to everything. "You need a machine to really understand content and understand people and be able to hold all that data," LeCun said. "That is an AI-complete problem."

Advertisement

Do we really need a machine to understand us? This sort of sounds like Stephen Hawking's worst nightmare, where AI destroys human kind as we know it. But on a more basic level, do you really want Facebook to censor you?