New Facebook blog series explains how they identify false news

MANILA, Philippines – Facebook has published the first entry in a new blog called “Hunt for False News” which is part of the social media platform’s effort to be more transparent to its users about misinformation.

The post, published Friday, October 19 by Facebook product manager Antonia Woodford examines 3 different fake stories that recently circulated on the site. Two of the stories were debunked by Facebook and third-party fact checkers but they missed the third.

Woodford in the blog shares their findings and how they identify content that are likely to be false. She shares different insights and methods of how they go about their fact-checking process.

For instance, the first story features video footage of a man in a headscarf spitting on a woman. The video was reportedly shared by multiple accounts and was accompanied by the caption “Man from Saudi spits in the face of the poor receptionist at a Hospital in London then attacks other staff.”

While the video, according to Agence France-Presse, is real and actually happened in Kuwait last year, it was recirculated and taken out of context with a falsified caption.

Woodford explains that Facebook has two primary methods to identify potentially false stories: they either use machine learning that takes into account user feedback of circulating news stories or ask help from third-party fact checking partners. In the case of the first story mentioned, it was the latter.

Facebook, meanwhile, also shared that new research claims that the volume of fake news on the social media site has been on a downward trend, and that interactions with false news sites have declined more than half since the 2016 US election.  

Woodford admits that Facebook sometimes miss certain stories that quickly circulate in the site. She assures users that they are already developing new technology to catch these kinds of stories and immediately demote them from users’ News Feeds. – Rappler.com