An Ugly Truth: Inside Facebook's Battle for Domination [calibre 4.99.5]

“It was obvious that our misinformation problem had become completely domestic,” said Stamos, who noted that in the weeks and months leading up to the vote, an array of websites and media companies had been successfully seeding false narratives about the vote through social media. Over the course of the day, Stamos’s team would file hundreds of reports to Facebook, Twitter, YouTube, and other social media companies about misinformation it had found. Of all the companies, Facebook was the fastest to respond and remove problematic content.

Still, rumors swirled online that voting machines were changing people’s ballots from Trump to Biden, that election officials were allowing people to vote under the names of their pet cats and dogs (or deceased relatives), and that certain states were purposefully miscounting votes. The rumors were baseless, but spread by President Trump and his supporters, they started to foster a narrative online that the election had, potentially, been stolen. “On Election Day and the day after, we saw a concerted effort by the president and his supporters in a way that was much more successful than what the Russians did in 2016,” noted Stamos. “In a way, it was beyond what the Russians could have hoped for.”

Facebook employees felt like they were fighting against an impossible tide of misinformation. They were being told to apply labels to any posts containing false information about the vote. But the labels, many of which simply directed people to Facebook’s Voting Information Center, were ineffective. Few people read them or stopped sharing content that was flagged, the data showed. “Imagine if the tobacco industry put a label on cigarettes which said, ‘Health officials have opinions about cigarettes. Go read this pamphlet about it!’ That is the equivalent of what we were doing. We weren’t coming out and saying, ‘This is a lie. There is no evidence for this. Here is the truth,’” a member of Facebook’s security team pointed out. “Or maybe, after years of allowing Trump to say whatever he wanted to say, people just got used to seeing crazy stuff on Facebook.”

As the night wore on, and it became clear that the election would be too close to call, Facebook’s election team realized its work was not done. Some of the remaining states could take days or even weeks to declare a winner. It was a scenario the team had prepared for, but had hoped would not happen. The day after the election, the team held a virtual meeting with Zuckerberg in which it requested he approve new break-glass emergency measures. These included an emergency change to the News Feed algorithm, which would put more emphasis on the “news ecosystem quality” scores, or NEQs. This was a secret internal ranking Facebook assigned to news publishers based on signals about the quality of their journalism. On one side of the spectrum were the outlets Facebook ranked as most trustworthy, such as the New York Times or the Wall Street Journal. On the other were outlets such as the Daily Caller, which had repeatedly been cited by Facebook’s fact-checkers for publishing false or misleading stories. Under the change, Facebook would tweak its algorithms so people saw more content from the higher-scoring news outlets than from those that promoted false news or conspiracies about the election results.