The past six months of headlines about Facebook’s partnership with fact-checkers paints a pretty bleak picture.
Bloomberg: “How 11 People Are Trying to Stop Fake News in the World’s Largest Election.”
The Guardian: “Facebook fact-checking in disarray as journalists push to cut ties.”
The Wall Street Journal: “In Facebook’s Effort to Fight Fake News, Human Fact-Checkers Struggle to Keep Up.”
It’s true: Only a few dozen fact-checking sites are working to limit the reach of false articles, photos and videos on Facebook. And, according to a Poynter survey from last summer, the staffs that work at those outlets frequently constitute fewer than five people.
But over the past year, Facebook has quadrupled its fact-checking partners — and it’s adding more in preparation for elections around the world. (Disclosure: Being a signatory of the International Fact-Checking Network’s code of principles is a necessary condition for joining the project.)
In a blog post published Thursday, product manager Antonia Woodford announced that Facebook added five new partners to its fact-checking program in the European Union ahead of next month’s parliamentary elections. Those outlets include: Ellinika Hoaxes in Greece, FactCheckNI in Northern Ireland, Faktograf in Croatia, Observador in Portugal and Patikrinta 15min in Lithuania.
“These organizations will review and rate the accuracy of content on Facebook,” Woodford wrote. “Our program now includes 21 partners fact-checking content in 14 European languages.”
In addition to debunking false content on Facebook, the new additions are contributing to FactCheckEU, a collaboration of 19 different news outlets fact-checking claims about the EU elections. Allan Leonard, editor of FactCheckNI, told Poynter in an email that the initiative was the main reason why he started working with Facebook.
“We were contacted by Facebook, on the back of our participation in the FactCheckEU project,” he said. “We are keen to be involved in any initiative that seeks to genuinely and comprehensively challenge misinformation and increases our impact.”
Thursday’s announcement brings Facebook’s total partners to 52 fact-checking sites in 33 countries worldwide (not counting AFP Arabic, which covers the Middle East and North Africa). That’s four times as many partners as April of last year, when the tech company was working with 13 fact-checkers to limit the spread of misinformation.
Under the partnership, once a fact-checker rates a post as false, its future reach in the News Feed is decreased and a fact check is appended. If users try to share it, they will receive a notification that it has been debunked. Additionally, pages that repeatedly share false posts have their monetization capabilities restricted.
The project first launched in December 2016 with five American fact-checkers: (Poynter-owned) PolitiFact, Factcheck.org, Snopes, ABC News and the Associated Press. Since then, it has grown at a steady clip, with the program doubling in size in just one month last spring. It has also expanded to limit the reach of false photos and videos in addition to articles.
With seven, India has the highest concentration of Facebook’s fact-checking partners. It’s followed closely by the United States and Indonesia, which both have six fact-checking outlets working with the company, and France, which has five.
In general, those numbers also reflect where the majority of Facebook’s users are. With more than 250 million users, India is the company’s biggest market, followed by the U.S., Brazil, Indonesia and Mexico, according to Statista.
Over the past year and a half, Facebook’s fact-checking program has become a central component of the company’s anti-misinformation efforts. Both CEO Mark Zuckerberg and COO Sheryl Sandberg have mentioned it in Congressional testimony.
And in an email to Poynter, Facebook said that it plans to expand the project even more in 2019. But growth has a limit; the IFCN only had 65 signatories of its code of principles as of publication. That means Facebook can only draw from that pool for future partners, leaving it 13 more potential projects to work with.
But the company isn’t concerned about scaling up its efforts.
“Expansion isn’t just about adding more total volume — it can also include expanding coverage where we already are,” said spokeswoman Lauren Svensson in an email. “Even though IFCN has an upper limit, we’re looking for ways to deepen our relationships with existing partners — for example, by working with them to cover additional languages and countries.”
Among Facebook’s fact-checking partners, there are some single news organizations that have fact-checking operations in multiple countries. The Agence France-Presse is the most prolific of the bunch, operating in 18 countries (not including its Arabic service), followed by Africa Check with four.
It’s notable that Facebook has essentially contracted misinformation moderation work to more than 50 independent media outlets around the world. But is it actually working?
In December, Poynter conducted a survey of 19 fact-checking organizations partnering with Facebook. We found that the majority of them said they thought the program has helped them reduce the reach of hoaxes on the platform. Nearly half of those surveyed said they had flagged more than 500 false posts since joining the partnership.
I write a weekly column detailing how some of the most popular fact checks perform on Facebook compared to the hoaxes they debunk. But more concrete data is still elusive.
In the past two-and-a-half years of Facebook’s fact-checking program, the only public numbers regarding its efficacy are contained in a leaked email obtained by BuzzFeed News in fall 2017. According to the email, the future reach of posts flagged as false decreases by up to 80% after an average of three days.
However, that may change soon.
Facebook announced on Monday which researchers it is partnering with to study the spread of misinformation on its platform. And in recent months, the company has started sharing personalized data reports with fact-checkers that illustrate the impact of their work. Some early research on the volume of misinformation on Facebook had some positive results.
Svensson said Facebook is working to improve the dashboard where fact-checking partners see what content users have flagged as potentially false. In the past, fact-checkers have criticized the tool for surfacing content that cannot be fact-checked, such as opinion posts and random photos and memes.
Facebook also is trying to amplify fact-checkers’ work using machine learning to identify duplicate claims and decrease the reach of pages that repeatedly share misinformation.
“Overall, we’re trying to shift to being more ‘claims focused’ so fact-checkers can focus on debunking net new claims (which we can then match to content) rather than requiring them to rate lots of content using existing fact checks,” Svensson said. “Another area we’re focused on is making sure fact-checkers’ work can have impact across surfaces.”
For the company’s new fact-checking partners, that’s still where the biggest benefits are.
“Working with Facebook provides the opportunity to directly inject our fact-checking remedy into this social network,” FactCheckNI wrote in a press statement Thursday. “Our aim is to improve the quality of information being read on Facebook and to increase the impact of our work.”
