Facebook’s DeepText AI Reads Your Posts and Chats Dylan Duarte June 1, 2016 News A few days ago I wrote a post about how robots are slowly but surely taking our jobs. Now we know they’re reading our social media messages, and they actually understand a good deal of what we’re talking about thanks to DeepText, Facebook’s new artificial intelligence system. It could be worse. They could have called it SkyNet for a laugh. Facebook sees 400,000 new stories and 125,000 comments every minute, so the job of the AI is to weed out the stuff users don’t want to see so that Facebook can promote the content they do want to see. Granted, everybody is different with different tastes, but nobody needs to read fifty different “congratulations” comments on your friend’s wedding pictures. That’s some of the stuff that DeepText will help hide. It just gets creepier from there, though. If you post on Facebook (or messenger) asking for a ride, Facebook will ask if you want an Uber. If you post about something for sale, Facebook will automatically format the post to suit your needs (okay maybe that one isn’t so creepy). Though just to assure you that this is indeed creepy, take note that I did say “or messenger” because yes, DeepText is reading your private messages. Here’s the thing, though. All of this makes Facebook more bearable. I don’t just mean that Facebook is hiding redudant and uninteresting posts, I mean that DeepText can catch harrassment, too, filtering out hate speech and abuse. It’s definitely a freedom versus security thing, but if it makes social media a more pleasant experience for those who suffer persecution, maybe it’s worth it. More importantly, though, it’s a robot apocalypse thing. Share This With The World!
