The numbers Facebook is and is not willing to share say a lot about its ability to stop election misinformation.
On a call with press Thursday, Guy Rosen, Facebook’s VP of Integrity, said that since March the company put “warning labels” on 180 million pieces of content — meaning a third-party fact-checker reviewed and debunked them.
Facebook subsequently put a gray box over that content when people shared it, with a “false” warning label and a link to the debunk.
That’s a big number: 180 million. Unfortunately, it doesn’t tell the whole story.
Facebook’s approach to all posts about the election was to put a far less obtrusive label at the bottom of them with a link to Facebook’s voting information center. The hub contains “authoritative information,” such as updates from secretaries of state, fact-checks, and articles from reputable news sources. It labeled posts claiming premature victory with a warning label saying that the election hadn’t been called yet.
More recently, it has been adding labels saying that Joe Biden is the projected winner of the election, with a link to the voting information center. Trump posts claiming a “rigged election” were labeled with, “The US has laws, procedures, and established institutions to ensure the integrity of our elections,” with a link to the info center.
We are continuing to label all posts from both presidential candidates making it clear that votes are still being counted and a winner has not been projected. We are also applying these labels to other individuals who declare premature victory in individual states or overall. pic.twitter.com/gmGdn4q52s
— Facebook Newsroom (@fbnewsroom) November 4, 2020
Facebook did not share how widespread the need for those labels was, nor did it respond to multiple questions from reporters about how effective the labels were at deterring the liking or sharing of those posts.
The elephant in the room was a recent article from BuzzFeed News reporting that Facebook employees internally knew that the labels aren’t that effective at stemming the tide of Trump’s lies.
”We have evidence that applying these informs to posts decreases their reshares by ~8%,” a Facebook data scientist told BuzzFeed. “However given that Trump has SO many shares on any given post, the decrease is not going to change shares by orders of magnitude.”
When asked about the efficacy of election labels, Rosen explained that the purpose of the labels was to provide more context and accurate information, and declined to comment on specific numbers. He told listeners that data points shared with press are highly vetted, and to not pay attention to other figures they might have seen. (Hmmm, wonder what he could be talking about?)
The company also said it removed “265,000 pieces of content from Facebook and Instagram in the US for violating our voter interference policies,” which means misinformation about how to vote, or posts meant to deter people from voting. It also rejected ad submissions 3.3 million times from unauthorized parties trying to run ads on social or political issues.
Meanwhile, the longer the election takes to certify, the more time Trump and his supporters have to beat their unsubstantiated drum about a “rigged election.” Much of that is happening on Facebook. Users keep making “Stop the Steal” groups. According to CrowdTangle data, all of the top politics post on Facebook over the past week come from Trump, most of which are spreading misinformation and undermining electoral integrity.
Thank goodness most of them come with a label at the bottom.
WATCH: How to recognize and avoid fake news
