December 5, 2016
After being heavily criticized for the spread of fake news during the latest U.S. presidential election, Facebook is now looking at the possibility of using its research in artificial intelligence to gain control over the problem. Facebook has done research in AI since late 2013, when it hired its current director of artificial intelligence Yann LeCun. But the company is moving gingerly into the field, still trying to figure out the pitfalls of AI, and how to introduce it sensibly and responsibly.
The Wall Street Journal reports that LeCun has said that AI technology “could be used to help stamp out fake news or detect violence in live videos by filtering the content on the site.” But the company is mulling over its use.
“What’s the trade-off between filtering and censorship? Freedom of experience and decency?” asks LeCun. “The technology either exists or can be developed. But then the question is how does it make sense to deploy it? And this isn’t my department.”
Still, LeCun admits that AI is “integral to the company’s operations,” and Facebook just released six informational videos about it. Going after fake news, says WSJ, poses the challenge of “discerning fact from fiction,” with the potential of “removing too much content with an AI filter.”
Two weeks ago, Facebook chief Mark Zuckerberg described that the company will build “systems to detect fake stories before users flag them, which would involve AI,” but, notes WSJ, “AI isn’t a panacea,” since, for example, “it doesn’t catch all terrorist propaganda.” Facebook director of applied machine learning Joaquin Candela reports that policing live video “requires a very fast computer vision algorithm,” which is in reach of his team, and also “developing a clear set of practices.”
In Recode, journalist Walter Mossberg takes a harder line, saying Facebook is not taking the fake news problem seriously, especially in light of the Pew Research Center’s finding that “roughly 44 percent of the U.S. adult population got at least some of its news from Facebook.” Mossberg suggests that Zuckerberg “stop pretending” the company isn’t a media company, and called Zuckerberg’s assertion that fake news didn’t impact the election as a “weaselly excuse.”
“I am also convinced that Facebook has the financial, technical and human resources to ferret out and totally block almost all fake news and hate speech, both of which it says it wants gone from its service,” Mossberg said, noting the company, which earned $3 billion in the last quarter, is “reportedly building a tool to prevent controversial content from appearing in News Feeds in countries like China.”
Facebook Working On a Plan to Pick News From Favored Media Partners Like Snapchat, Business Insider, 12/2/16