Facebook’s ‘trustworthiness’ scores weed out fake news, false user reports | Social
Facebook probably has a scorecard for you — in a recent interview with the Washington Post, Facebook product manager Tessa Lyons shared that the network has a system for determining which users are more trustworthy. The “score” is used to push potential fake news to the top of the list for fact-checking organizations.
Lyons explains that the score works to classify users that have reported posts on Facebook using the reporting tool. Some users, she says, report posts not because they are false but because they don’t agree with the content. Users that regularly flag posts that actually contain fake news will have a high score. Users that flag posts that waste fact-checker time have a lower score.
By factoring in the “trustworthiness” of the user, those user-reported fakes are arranged in a way that helps make the best use of fact-checker time, the company says. The score is used along with other factors that help the fact-checking team determine which posts to look at first.
Facebook isn’t going to tell you what your score is, however, or even tell you how that score is generated. The network is keeping the scoring system confidential under the belief that bad actors could use that information to artificially boost their trustworthiness score. Facebook also didn’t say if every user has a score or not and also didn’t divulge if there are other ways the network uses those scores.
The score was developed sometime over the past year, Lyons said.
The interview offers a glimpse into what’s going on behind the scenes as Facebook works to clean up the platform, or more accurately, to attempt to prevent false information from going viral. The company has launched a host of new initiatives to try to curb the spread of fake news, with varying degrees of success. The network also ranks the trustworthiness of publications using user opinion instead of attempting the task themselves.
The network now uses related articles to attempt to shed more light on the subject, along with including more information about publishers. One feature uses Wikipedia to crowd-source related information. The company has also launched partnerships with organizations such as the Atlantic Council to improve the platform’s approach to fake news.