Facebook Plots an AI that Will Tell You Not to Post Drunk Selfies
Almost everyone who has a Facebook has posted a picture that they've ultimately regretted. Out of the goodness of their hearts (I guess), Facebook wants to help you put a stop to this potentially career/reputation-damaging behavior and create an artificial intelligence that will prevent you from posting pictures of your intoxicated self, or at least give you fair warning before you go ahead and do it anyway.
Yann LeCun, head of the Facebook Artificial Intelligence Research lab (that name really sounds like they're trying to take over the world, doesn't it?), is working on building a Facebook digital assistant that will, among other things, let you know when you're posting a picture that you'll be embarrassed about in the morning. This assistant will "virtually tap you on the shoulder and say, 'Uh, this is being posted publicly. Are you sure you want your boss and your mother to see this?'" LeCun told Wired.
Since the project is in its infancy, it's difficult to know how helpful it would be. But if it would just warn you that you might be doing something regrettable, then its helpfulness might be somewhat limited, as many people would just say, "Yes, I'm sure," and post the pictures, maybe even throwing some profanity at the assistant in the process. And it seems unlikely that it would prevent you from uploading the pictures entirely, as it presumably can't know for certain that you're drunk (unless it teams up with Oxford's eDNA, that is).
But on the other hand, knowing enough about you that it can know for a fact that you're drunk might be the logical endpoint of this technology. Facebook can already recognize your face and other friends' faces in your pictures and analyze your online interests in order to give you relevant content in your Newsfeed. LeCun envisions this digital assistant as part of a larger "deep-learning" process that takes it to the next level, and allows one intelligence to control all of your information on Facebook.
"Imagine that you had an intelligent digital assistant which would mediate your interaction with your friends, and also with content on Facebook," said LeCun. "You will have a single point of contact to mediate your interaction but also to protect your private information."
This means that the assistant would be better able to collate all of your personal information in order to help other companies target you for advertising, but it would also mean that it knows enough about you to allow you to control your online persona. For example, if the face recognition technology was a little more advanced, then it could warn you when anyone on Facebook, even someone without any direct connection to you, posts a picture of you without your consent. Like Oxford's eDNA, it's a mixed bag, in which more advanced knowledge of your person allows both higher security and more privacy violations by the website itself.