After months of debate and disagreement over the handling of inflammatory or misleading posts from Donald Trump, Facebook employees want CEO Mark Zuckerberg to explain what the company would do if the leader of the free world uses the social network to undermine the results of the 2020 US presidential election.
“I do think we’re headed for a problematic scenario where Facebook is going to be used to aggressively undermine the legitimacy of the US elections, in a way that has never been possible in history,” one Facebook employee wrote in a group on Workplace, the company’s internal communication platform, earlier this week.
For the past week, this scenario has been a topic of heated discussion inside Facebook and was a top question for its leader. Some 2,900 employees asked Zuckerberg to address it publicly during a company-wide meeting on Thursday, which he partly did, calling it “an unprecedented position.”
Zuckerberg’s remarks came amid growing internal concerns about the company’s competence in handling misinformation, and the precautions it is taking to ensure its platform isn’t used to disrupt or mislead ahead of the US presidential election. Though Facebook says it has committed more money and resources to avoid repeating its failures during the 2016 election, some employees believe it isn’t enough. President Trump has already spent months raising questions about the legitimacy of the upcoming 2020 election, spreading misinformation about mail-in ballots, and declining to say if he’d accept the possibility of losing to Democratic nominee Joe Biden in November.
In July, Trump told Fox News he wasn’t sure if he’d concede to Biden, casting doubt on whether there would be a peaceful transition of power if the former vice president wins the election. “I have to see. I’m not just going to say yes. I’m not going to say no,” the president said.
“America can’t afford for Facebook to take a wait-and-see approach when it comes to the integrity of our democracy.”
On Facebook’s internal message boards, discussion about the Trump election question remained civil prior to Thursday’s all-hands meeting. Employees debated the merits of censoring a sitting president’s potentially false statements about election results with one person noting that “it would be a really troubling policy to apply globally.”
“America can’t afford for Facebook to take a wait-and-see approach when it comes to the integrity of our democracy,” said Jesse Lehrich, a former foreign policy spokesperson to Hillary Clinton and cofounder of Accountable Tech, a nonprofit advocacy group. “Unless they proactively outline clear policies and enforcement mechanisms to safeguard the election, the platform will be weaponized to undermine it.”
On Thursday, Zuckerberg told employees that the increased use of mail-in ballots due to the pandemic will likely lead to a situation where election results will not be available “for days” or “for weeks.” He noted political figures and commentators may attempt to try to call an election early, in which case the company may label a post explaining that results are not yet final.
Zuckerberg did not have a clear answer for what the company would do should Trump declared the election results invalid.
“This is where we’re in unprecedented territory with the president saying some of the things that he’s saying that I find quite troubling,” he said. “We’re thinking through what policy may be appropriate here. This is obviously going to be a sensitive thing to work through.”
While there are signs Facebook will stand up to Trump in cases where he violates its rules — as on Wednesday when it removed a video post from the president in which he claimed that children are “almost immune” to COVID-19 — there are others who suggest the company is caving to critical voices on the right. In another recent Workplace post, a senior engineer collected internal evidence that showed Facebook was giving preferential treatment to prominent conservative accounts to help them remove fact-checks from their content.
The company responded by removing his post and restricting internal access to the information he cited. On Wednesday the engineer was fired, according to internal posts seen by BuzzFeed News.
“Intervening in Fact-Checks”
With heightened internal tensions and morale at a low point, concerns about how the company handles fact-checked content have exploded in an internal Workplace group dedicated to misinformation policy.
Last Friday, at another all-hands meeting, employees asked Zuckerberg how right-wing publication Breitbart News could remain a Facebook News partner after sharing a video that promoted unproven treatments and said masks were unnecessary to combat the novel coronavirus. The video racked up 14 million views in six hours before it was removed from Breitbart’s page, though other accounts continued to share it.
Zuckerberg danced around the question but did note that Breitbart could be removed from the company’s news tab if it were to receive two strikes for publishing misinformation within 90 days of each other. (Facebook News partners, which include dozens of publications such as BuzzFeed News and the Washington Post, receive compensation and placement in a special news tab on the social network.)
“This was certainly one strike against them for misinformation, but they don’t have others in the last 90 days,” Zuckerberg said. “So by the policies that we have, which by the way I think are generally pretty reasonable on this, it doesn’t make sense to remove them.”
But some of Facebook’s own employees gathered evidence they say shows Breitbart — along with other right-wing outlets and figures including Turning Point USA founder Charlie Kirk, Trump supporters Diamond and Silk, and conservative video production nonprofit Prager University — has received special treatment that helped it avoid running afoul of company policy. They see it as part of a pattern of preferential treatment for right-wing publishers and pages, many of which have alleged that the social network is biased against conservatives.
“We defer to third-party fact-checkers on the rating that a piece of content receives,” Facebook spokesperson Liz Bourgeois said in a statement. “When a fact checker applies a rating, we apply a label and demotion. But we are responsible for how we manage our internal systems for repeat offenders. We apply additional system wide penalties for multiple false ratings, including demonetization and the inability to advertise, unless we determine that one or more of those ratings does not warrant additional consequences.”
On July 22, a Facebook employee posted a message to the company’s internal misinformation policy group noting that some misinformation strikes against Breitbart had been cleared by someone at Facebook seemingly acting on the publication’s behalf.
“A Breitbart escalation marked ‘urgent: end of day’ was resolved on the same day, with all misinformation strikes against Breitbart’s page and against their domain cleared without explanation,” the employee wrote.
The same employee said a partly false rating applied to an Instagram post from Charlie Kirk was flagged for “priority” escalation by Joel Kaplan, the company’s vice president of global public policy. Kaplan once served in George W. Bush’s administration and drew criticism for publicly supporting Brett Kavanaugh’s controversial nomination to the Supreme Court.
Aaron Sharockman, the executive director of PolitiFact, told BuzzFeed News a contact at Facebook did call to discuss Kirk’s post.
“We had a call with them where they wanted to know how this post was aligned with the program,” Sharockman said. “Was this just a minor inaccuracy or was it something that we thought was something that had potential harmful effects?”
PolitiFact did not change its rating on the post. “We stuck to our guns there,” Sharockman said.
Past Facebook employees, including Yaël Eisenstat, Facebook’s former global election ads integrity lead, have expressed concerns with Kaplan’s influence over content enforcement decisions. She previously told BuzzFeed News a member of Kaplan’s Washington policy team attempted to influence ad enforcement decisions for an ad placed by a conservative organization.
Facebook did not respond to questions about why Kaplan would personally intervene in matters like this.
“It appears that policy people have been intervening in fact-checks on behalf of *exclusively* right-wing publishers.”
These and other interventions appear to be in violation of Facebook’s official policy, which requires publishers wishing to dispute a fact check rating to contact the Facebook fact-checking partner responsible.
“It appears that policy people have been intervening in fact-checks on behalf of *exclusively* right-wing publishers, to avoid them getting repeat-offender status,” wrote another employee in the company’s internal “misinformation policy” discussion group.
Individuals that spoke out about the apparent special treatment of right-wing pages have also faced consequences. In one case, a senior Facebook engineer collected multiple instances of conservative figures receiving unique help from Facebook employees, including those on the policy team, to remove fact-checks on their content. His July post was removed because it violated the company’s “respectful communication policy.”
After the engineer’s post was removed, the related internal “tasks” he’d cited as examples of the alleged special treatment were made private and inaccessible to employees, according to a Workplace post from another employee.
“Personally this makes me so angry and ashamed of this company,” wrote the employee in support of their colleague.
The engineer joined the company in 2016 and most recently worked on Instagram. He left the company on Wednesday. One employee on an internal thread seen by BuzzFeed News said that they received permission from the engineer to say that the dismissal “was not voluntary.”
In an internal post before his dismissal, the engineer said he was “told to expect to be contacted by legal and HR if my post is found to be violating other policies in addition to Respectful Comms.”
Facebook denied the employee had been terminated for the post but said it was because “they broke the company’s rules.”
“We have an open culture and encourage employees to speak out about concerns they have,” Bourgeois said.
News of his firing caused some Facebook employees to say that they now fear speaking critically about the company in internal discussions. One person said they were deleting old posts and comments, while another said this was “hardly the first time the respectful workplace guidelines have been used to snipe a prominent critic of company policies/ethics.”
“[He] was a conscience of this company, and a tireless voice for us doing the right thing,” said another employee.
The terminated employee declined to comment and asked not to be named for fear of repercussions.
The internal evidence gathered by the engineer aligns with the experience of a journalist who works for one of Facebook’s US fact-checking partners. They told BuzzFeed News that conservative pages often complain directly to the company.
“Of the publishers that don’t follow the procedure, it seems to be mostly ones on the right. Instead of appealing to the fact-checker they immediately call their rep at Facebook,” said the journalist, who declined to be named because they were not authorized to speak publicly. “They jump straight up and say ‘censorship, First Amendment, freedom.’”
“I think Facebook is a bit afraid of them because of the Trump administration,” they added.
“I think Facebook is a bit afraid of them because of the Trump administration.”
Facebook typically assigns dedicated partner managers to pages with large followings or big ad budgets. They help their clients maximize their use of the platform. But in the cases identified in the engineer’s post, partner reps appear to have sought preferential treatment for right-wing publishers. This resulted in phone calls to fact-checking partners from people at Facebook, and instances where misinformation strikes appear to have been removed from content without a fact-checker’s knowledge or involvement.
A Facebook employee’s July 22 post restating the engineer’s findings identified multiple cases in which a fact-check complaint from a right-wing page was escalated and in some cases resolved in the account’s favor the same day.
According to the post, Joel Kaplan flagged a fact check of a Charlie Kirk Instagram post for resolution “ASAP/before 12 p.m. ET”. This same employee claimed PragerU’s partner manager was part of a “two weeks long effort” to prevent the site from being given Repeat Offender status, a designation that would have limited its reach and advertising privileges.
Citing “partner sensitivity,” the rep noted that PragerU runs a lot of ads, and argued that the content in question qualified as opinion and was therefore exempt from being fact-checked.
Facebook did not answer questions about why a partner manager would cite ad volume as a reason for not acting against a group of pages.
Sharockman said PolitiFact’s contacts at Facebook have never asked them to change a rating. But Facebook reps do reach out to discuss whether a post is opinion or otherwise outside of the scope of the program.
“We have had discussions where our partners have asked us, ‘Why did you fact check this? Why did you come to this conclusion? How do you think it fits within the scope of Facebook’s rules or regulations within fact-checking?’” he said.
“We sometimes reach out to fact-checkers to clarify our guidelines and scope of the program,” Bourgeois said.
Mark Provost manages multiple large progressive Facebook pages, including The Other 98%, one of the biggest on the platform. He said his Facebook partner manager is far less responsive than what the Facebook employee documented for Provost’s counterparts on the right. And he’s not aware of any case where Facebook contacted a fact-checker on his behalf.
“We don’t get a message back for 10 days,” Provost said of Facebook. “I imagine the right wing is getting a way better deal.”
His Other 98% page is currently at risk of deletion because it has three strikes. Provost said one strike is for a hate speech violation three months ago, though as of now no one at Facebook has told him what the offending post was. His partner representative said they would look into it, according to Provost. That was last week and he says he still hasn’t received an update.
“I’m getting so frustrated with this,” Provost said. “The best solution would be if Facebook were as responsive as the fact-checking companies that they’ve assigned.”
Provost said Facebook’s fact-checking partners are easy to deal with when he contacts them to dispute a rating. In some cases, he provided additional proof of his claim to get a rating changed; in other cases, he corrected the offending post and the false rating was removed.
The fact-checker who spoke to BuzzFeed News said most page owners follow the policy and contact them to dispute a rating. But in some cases, they hear directly from the Facebook partner manager assigned to work with fact-checkers.
“They will ask us, ‘Could you take a look at this again? Are you sure?’” the checker said.
In other cases, Facebook itself will quietly remove a fact-check applied by one of its partners. That appears to be what happened with a March 25 post from Diamond and Silk. The duo wrote on Facebook, “How the hell is allocating 25 million dollars in order to give a raise to house members, that don’t give a damn about Americans, going to help stimulate America’s economy? Tell me how? #PutAmericansBackToWorkNow.”
Lead Stories, a Facebook fact-checking partner, rated the post false. Diamond and Silk initially followed the proper procedure and appealed directly to Lead Stories.
“They were detailed in their appeal and we replied promptly,” Alan Duke, the editor-in-chief of Lead Stories, told BuzzFeed News. “The result was that we decided that a false rating initially given their content should be revised downward to ‘partly false.’”
But that partly false rating apparently didn’t satisfy Diamond and Silk, who reportedly lost their online Fox News show after spreading coronavirus misinformation. As detailed in the July 22 internal Facebook post, Diamond and Silk appealed to their partner manager, who opened an internal ticket for the issue. The Facebook rep argued the post was opinion and warned that the duo “is extremely sensitive and has not hesitated going public about their concerns around alleged conservative bias on Facebook.”
While the partly false fact-check is still visible on their post, employees said Facebook removed the strike against their account internally.
The Facebook employee said “someone described on the task as ‘Policy/Leadership’ made the call to not only completely remove this strike, but also the one from January.”
PolitiFact’s Sharockman said he doesn’t know what goes on inside Facebook when it comes to applying or removing strikes for misinformation. All the Facebook fact-checking partners can do is try to be honest, even if the other parties involved aren’t.
“We’re trying to be intellectually honest and thoughtful and deliberate and open and transparent. And no one else is. Everyone else is using the system to their personal benefit.” ●
Correction: The Lead Stories fact-check for a March post from Diamond and Silk is displayed in the “Related Articles” section of their post. This story incorrectly said it had been removed.