This week, Facebook CEO Mark Zuckerberg announced that something resembling a Facebook Dislike button is on its way. Finally.
"People have asked about the dislike button for many years. Today is a special day because today is the day I can say we’re working on it and shipping it," Zuckerberg said during a Q&A at Facebook headquarters in Menlo Park Tuesday.
Facebook isn't adding the feature just because we've been begging for it pretty much since the inception of the Like button in 2009. It's another way to feed data into the algorithms that Facebook uses to serve you ads and content, the stuff that makes the company $$$.
From a machine intelligence perspective, the dislike button is going to give Facebook more insight into user preferences, said Chris Nicholson, CEO of AI company Skymind. "It adds a new dimension to what they know about us," he added.
Right now, when we interact with a person or brand using the Like button, Facebook's algorithms read that as our wanting to see more of that kind of content. It's an easy signal. The button that signals dislike—whatever Facebook decides to call it—will not be as clear a signal, because there are so many reasons we might signal dislike. For example, you dislike every post about the Red Sox because you hate baseball, or because you're a Yankees fan. A big challenge for machines is dealing with ambiguity.
"Arguably, it’s harder to infer the true meaning of a dislike on a post or page than a like," said Aditya Jami, an infrastructure engineer who previously built large-scale computing systems at Netflix and Yahoo. "For example, someone can dislike a sports team—and any positive content associated with it—just because he or she is an avid supporter of the rival team when compared to hitting dislike for expressing empathy towards disaster news."
It's likely that Facebook will also look to what people type in the comment boxes to assess how they feel about content, their friends and brands they interact with. In recent years, the social network has been beefing up its investments in artificial intelligence, particularly in image recognition, but also in natural language processing, the science of understanding written language. The more you write, the more data its team of AI experts have to train systems to understand your wants and needs. Coupled, with a Like and Dislike feature, the social network will now have even more data on which to judge whether you're likely to want to interact with the ads, photos, videos and status updates that make it into your News Feed.
That's helpful not just to Facebook, but also to the ecosystem of companies built on top of it. Jami recently founded Predict Effect, a startup that uses data from blogs, news articles and social networks like Facebook and Twitter to create profiles of users that marketers can use to target ads and grow their user base. The Dislike button could be a boon for social intelligence businesses like this.
The company currently infers "dislike" using sentiment analysis of posts and comments, said Jami, but that's not as reliable a signal as a dislike button.
"We are extremely excited about this announcement since it will help us build richer interest signatures for audience buckets and thus able to extract more fine grained targeting to acquire new users for our customers," Jami said.
And here is where the new button will make things complicated: will you like this article because it was informative or dislike it because you don't like what it helped you learned about the Facebook world to come?
Daniela Hernandez is a senior writer at Fusion. She likes science, robots, pugs, and coffee.