SAN FRANCISCO - One of Facebook’s third-party fact-checking partners on Tuesday said the company needed to share more data with fact-checkers to better curb misinformation.
UK-based charity Full Fact, which joined the program in January, issued a report urging Facebook to provide more data on how flagged content is shared over time, to see how quickly false information is spreading and assess how fact-checks are containing the spread.
The charity, which said it had earned $171,800 from fact-checking Facebook content between January and June, also raised concerns about the effectiveness of Facebook’s machine-learning approaches for identifying potentially false content.
Facebook launched the program in December 2016 and now has 54 fact-checking partners working in 42 languages.
Under the system, fact-checking partners rate content with labels and Facebook then downgrades the distribution of certain items. Facebook also adds fact-checkers’ articles to a story’s “Related Articles” and notifies users if they try to share or have previously shared a false story.
Julia Bain of Facebook’s integrity partnerships team said the social media network was evolving its rating scale to account for different types of misinformation and working on improving the impact of fact-checks by identifying similar or identical content.
“We are encouraged that many of the recommendations in the report are being actively pursued by our teams,” Bain said in a statement to Reuters.
Although Full Fact said it still thought the program was worthwhile, it echoed statements from other partners that Facebook was not sharing enough data that they could use to evaluate their fact-checks.
Eugene Kiely, director of FactCheck.org, recently told Reuters that he “couldn’t make heads or tails” of a report the organization received last December.
“We haven’t really received anything from Facebook that would provide evidence one way or the other to the effectiveness,” he added.
Full Fact also called for the program to be expanded to Instagram, so that fact-checkers could directly check content on the platform. In May, the company started using image-detection to find content debunked on Facebook and removing it from Instagram’s Explore tab and hashtag search results.