Facebook struggles to purge fake news


2017-05-05 / www.ft.com

Accused of influencing the US election, the social media group is now using more fact-checkers


Tucked away on the first floor of a brightly painted backpackers’ hostel in central Berlin is Correctiv, a small non-profit media start-up with a big job: helping to save Facebook from an epidemic of fake news.

David Schraven, its proprietor and chief editor, is a bearded, balding former journalist who used to head up investigations at Funke Mediengruppe, Germany’s third-largest media publisher. His background is in investigating neo-Nazis and underground terror organisations, not technology and social networks. “I never cared about Facebook or its relationship to the media,” he says.

But his attitude changed after watching the impact of online “fake news” in the campaigns of Britain’s EU referendum and the US election. He began to worry that “disruptive forces” spreading deliberately false news reports could influence Germany’s national poll in September, when Chancellor Angela Merkel is standing for a fourth term.

Sitting in Correctiv’s tchotchke-filled office, Mr Schraven checks the facts of a story that has gone viral about refugees who allegedly raped a German woman and threw her out of a car window. It was originally published by a website called “Rape-fugees”.

Such stories have become more common recently in Germany, where Ms Merkel’s open-arms policy towards Syrian refugees — part of the more than 1m refugees the country has admitted since 2015 — remains a divisive political issue. “People are starting to say: we need to get these refugees out, we need a clean country. I’ve seen the mood change in my own town because of these news stories,” he says. “I think [fake news] could have a big impact on German elections.”

After confirming that the refugee story is untrue, he flags it to Facebook, which then alerts its users that it has been disputed by a fact-checker.

Correctiv’s fact-checking is just one part of Facebook’s evolving response to a series of controversies that have shaken the company over the past year. The social media group has been accused of influencing the US presidential election by turbo-charging the spread of fake news stories and creating “filter bubbles” that isolate voters from other opinions.

Facebook’s video platform has also been criticised after some grisly episodes, including the recent video of a murder in Ohio that stayed up for two hours after being flagged by users and a recording of a Thai man killing his baby daughter that took 24 hours to remove.

The outcry has raised deeper questions about the nature of Facebook and its social responsibilities. Throughout its 13-year rise to 1.9bn users and a commanding share of the digital advertising market, the group has maintained that it is a “neutral technology platform” with very limited responsibility for the content it hosts. Its only social obligation, it has maintained, is to be a conduit for connecting people.

But over the past year, it has been harder for the company to maintain this argument.

After President Barack Obama complained about the “dust cloud of nonsense” on the social network before the election, Mark Zuckerberg, Facebook founder and chief executive, appeared to play down the site’s power to influence politics. It was a “pretty crazy idea”, he said, to say that fake news articles on the site before the November election, such as one saying that Pope Francis had endorsed Donald Trump, had influenced the outcome.

His stance drew criticism, including from within the group’s Silicon Valley headquarters, and Mr Zuckerberg rethought his position. Facebook then launched several initiatives, including partnering with fact-checkers such as Correctiv, to try to stop the spread of fake news on the site.

…But critics say the company has not gone far enough to recognise a simple fact: it is a media company. As the world’s largest disseminator of news, they say, Facebook should adopt editorial standards that reflect its power.

“For a long time they have resisted actively being called a media organisation because it implies they might be regulated as one, and expected to generate public good,” says Philip Howard, professor of internet studies and director of research at the Oxford Internet Institute. “I think they have to be publicly audited and they should be taking some editorial role.”

…Facebook has resisted calls for hiring an editorial director or taking other steps that would make it more like a traditional news organisation. Instead, it is tweaking its technology to stop the spread of misinformation, and offering to pay fact-checkers to ensure the veracity of stories on the platform.

“A commercial relationship is something that’s on the table and that we are very open to,” Adam Mosseri, head of Facebook’s news feed, told the Financial Times last month. “It could depend on individual organisations, but we want to engage responsibly and if that means a financial arrangement, we are very open to it.”

In the weeks after the US election, the group appeared to take more responsibility for its role in shaping the news agenda. It began working with American fact-checkers such as Snopes and Politifact, later expanding to Germany and France, where it worked with Le Monde and other local media, ahead of their national elections. Mr Schraven’s partnership with Facebook Germany began in March, when Correctiv’s team began to trawl through the new fact-checking system.

The Correctiv system starts with users reporting a story as false. Once it is flagged, it appears on a list of links for Correctiv to check. Each story is ranked by how popular it is on Facebook; a barometer shows how many actions, comments, shares or likes each story has received to help fact-checkers prioritise their work. Every checked story has a flag attesting to its veracity. If it is false, a link to an alternative, factual version is included.

“You can still read both versions of the story but when you want to share it — and this is the most important thing — you will be warned that independent fact-checkers do not think it is trustworthy,” says Mr Schraven.

…But the more Facebook becomes involved in policing content on its services, the greater the risk that it will alienate users or potential advertisers.

“No one at Facebook has ever admitted they are a media company, because that will have an impact on how they are treated in US law,” says Prof Howard. “If they are considered a media company and are taking money for placement of advertisements, then they are responsible for truth in advertising and have to dedicate budget to public service ads, like all other media companies do.”

With the group’s profits reaching $10bn last year and the stock up 30 per cent this year, the recent controversies have had little impact on investor sentiment. “Investors are not necessarily social critics,” says Brian Wieser, an analyst at Pivotal research.

Facebook users spend an average of 50 minutes a day using its apps, Facebook, Instagram and Facebook Messenger. But what if efforts to expose people to other political views result in lower engagement? The group’s algorithm is designed to keep users on the site so it can show them more advertising. But that does not necessarily fit with its new aspirations to help readers find an accurate source of information.

“What [are] the commercial consequences? How much is he willing to sacrifice revenue in order to solve the problem? That is the fundamental question he does not address,” says Mr Kirkpatrick.

Ultimately, Facebook’s decisions on what to show people in their news feeds are opaque and, even with the best intentions, may not represent its amorphous constituency of 2bn users with different nationalities, ages and cultural norms…

In April, Facebook launched a $14m project aimed at improving the integrity of news online. It also launched an educational tool to help spot false news on Facebook, via a list of 10 tips. …






Attacking Hillary: Clinton remains a popular target

A story saying Pope Francis had endorsed Donald Trump is estimated to have been shared almost 1m times in the lead up to the US presidential election.

Now, six months after the vote, anti-Hillary Clinton stories continue to dominate the world of fake news.

Snopes, one of the oldest fact- checkers on the web, compiles a list of the most “intriguing and questionable items” making the rounds online.


Below are the top five fake stories last week, three of which feature Mrs Clinton:


Eddie Murphy death hoax An old story about comedian Eddie Murphy was recirculated after his brother Charlie died. The article was published by Linkbeef, which Snopes says is a “purveyor of fake news and celebrity death hoaxes”.

Clinton-Benghazi A story published by The Last Line of Defense claimed that Trey Gowdy, the US representative who chaired the select committee that questioned Hillary Clinton about the 2012 attack against US facilities in Benghazi in Libya, was put into protective custody after his investigators were found dead. Snopes found that the story was fake.

Obama mistress A story claiming Michelle Obama had filed for divorce over President Barack Obama’s pregnant mistress was also published by The Last Line of Defense. Snopes calls it “false”.

Clinton ally dead Was Mrs Clinton’s assistant JW McGill found dead? A frequent topic of fake news stories — that the Clintons purportedly killed for their political careers — re-emerged. Snopes rates the story as “false”.

Chelsea connection Was Chelsea Clinton arrested for having sex with a 13-year-old boy? Maga.news provided no information about the alleged arrest and there is nothing to back it up, says Snopes.