The world took notice on May 26, when Twitter fact-checked President Donald Trump for the very first time. Trump posted a series of blatant lies about mail-in voting, and declared that “this will be a rigged election.” Twitter responded swiftly, saying that the viral posts contained “potentially misleading” information, and slapped a fact-check label on them.
But seven weeks later, and after a dozen similarly untruthful tweets from the President, that extraordinary step by Twitter looks more like a one-time aberration than the new normal.
According to a CNN analysis, Trump’s latest posts included misleading information about the mechanics of mail-in voting, flat-out lies that the system is plagued by widespread fraud, and false accusations that Democrats are using these new voting procedures to cheat. He repeated the phrase “rigged election” five times. Yet, Twitter took no action.
A Twitter spokesperson told CNN that many of these posts are “currently not in violation of the Twitter Rules,” even though they struck a similar tone to the tweets they sanctioned in May.
The distinction Twitter is drawing is that there’s a difference between questioning the integrity of mail-in voting as a broad concept, versus suggesting that voting procedures in a particular state are fraudulent. The two posts Twitter fact-checked in May specifically called out California.
This approach creates a strange dynamic. Untrue tweets about vote-rigging in a specific state are unacceptable and subject to fact-checking. But Twitter says its hands are tied when those same lies are spread on a national scale, where they could influence millions of more voters.
By drawing this line in the sand, Twitter can say it is consistently applying its rules. At the very least, Twitter deserves credit for trying to hold public figures accountable. The same problem is pervasive on Facebook, which became a breeding ground for misinformation in recent years, and which has taken fewer steps than Twitter despite commanding a much larger audience.
Still, experts told CNN that Twitter’s narrow policy misses the bigger picture: Even if a tweet doesn’t mention a specific state, it can still have the same impact of undermining faith in US elections and suppressing turnout, which is precisely what Twitter says it is trying to prevent.
“I am sure that if I ran a randomized, controlled study and measured the impact of these tweets on public confidence in the legitimacy of the election, I can’t imagine I would find a distinction between those posts,” said professor Joshua Tucker, who is the co-director of the Center for Social Media and Politics at New York University. “If we’re actually worried about people’s belief in the sanctity of the American electoral process, those tweets don’t look any different to me.”
Drawing a fine line
Over the past year, Twitter has diligently refined its policies regarding manipulated media and “deepfake” videos, and announced a new standard for tweets from world leaders that contain false and potentially dangerous information. Twitter solicited input from thousands of users and concluded that most people want problematic posts to be hidden or labeled – not removed.
Twitter also rolled out a “civic integrity policy” in May. The rules say, in part, that users “may not use Twitter’s services for the purpose of manipulating or interfering in elections or other civic processes.” The company said this includes “posting or sharing content that may suppress participation or mislead people about when, where, or how to participate in a civic process.”
The rules draw a fine line. False claims about supposed poll closures, equipment problems or disruptions could dissuade people from voting, Twitter says, and therefore violate the rules.
But “broad, non-specific statements about the integrity of elections” are tolerated, according to the company’s written policy. Conveniently, in describing what doesn’t break the rules, Twitter explicitly carved out an exception to allow “unsubstantiated claims that an election is ‘rigged’” – which has been one of Trump’s go-to phrases for years, dating back to the 2016 campaign.
The company concedes that this still allows some “false or untrue information” to flow freely on its platform. It’s all part of the give-and-take of social media platforms patrolling political speech.
“There are an infinite number of ways to cast doubt about mail-in balloting,” said Stanford Law School professor Nathaniel Persily, an expert on how social media impacts democracy. “Some of them are clear voter suppression, some are disinformation, but some are predictions and opinions. It’s difficult for the tech companies to draw the line between these three categories.”
Twitter gets tested
Rigid rules like these give Trump an opening to spread his harmful messaging about fraudulent elections – unencumbered – even if his posts are filled with debunked allegations and rumors.
As the coronavirus pandemic swept through the country, Democratic and Republican election officials sprang into action to offer more mail-in voting options. In response, Trump repeatedly said expanded mail-in voting threatens his re-election and hurts Republicans across the board, even though nonpartisan experts say neither party automatically benefits from postal voting.
On May 26, Trump took aim at the largest Democratic state, which announced it would send mail ballots to all registered voters: “The Governor of California is sending Ballots to millions of people, anyone living in the state, no matter who they are or how they got there, will get one.”
At this time, Twitter was facing an avalanche of criticism for refusing to remove or fact-check Trump’s posts on a different topic. He had falsely accused TV host Joe Scarborough of murdering a staffer who died in his congressional office in 2001. The death was an accident and Scarborough wasn’t involved. Twitter took no action, even after pleas from the staffer’s widower.
Twitter decided that Trump’s election-related posts crossed the line and merited a response.
Twitter added fact-check labels to two of his tweets, urging users to “get the facts about mail-in ballots,” and linking to a curated page of articles debunking Trump’s claims. Despite this highly publicized wrist-slapping, Trump instantly returned to the rhetoric that got him in trouble.
Within days, Trump posted four tweets with similarly false claims. Twitter repeatedly declined to take action, saying he didn’t break the rules. Trump spread a new lie in late June, claiming that “MILLIONS OF MAIL-IN BALLOTS WILL BE PRINTED BY FOREIGN COUNTRIES.” According to Twitter, this fell into the category of “broad, non-specific statements” about election integrity.
On Friday, Trump falsely claimed, “mail-In Ballot fraud (is) found in many elections,” and raised the possibility of millions of illegal votes this fall, tweeting, “20% fraudulent ballots?” Again, Twitter demurred, pointing to the carveout that shields untrue rhetoric about “rigged” elections.
“Trump’s style of posing these questions at the end of his tweets, that is borrowed from the National Enquirer, in the way they will often avoid defamation lawsuits by simply raising the question,” Persily said. “And that is the problem. There is no way the platforms can win on this.”
The way forward
The back-and-forth over Trump’s posts pulls Twitter – and other platforms like Facebook – into a position they’ve resisted since their inception. Even though hundreds of millions of people get information from their global platforms, they’ve tried to avoid deciding what is true and false.
“It’s a legitimate concern about how much you want the platforms to be arbiters of truth about politics,” said Tucker, the NYU professor. Twitter shouldn’t face too much criticism for refusing to publicly rebuke every Trump falsehood about mail-in voting because “that’s not what Twitter’s policy is, they aren’t flagging any tweet that undermines confidence in the election process.”
Even though these companies relented, and are now refereeing some speech, they’ve been rebuked by both parties. Asked why Twitter’s policies toward election-related misinformation treat state-specific lies more harshly than lies about voting on a national level, a spokesman reiterated that the purpose of the policy is to uphold civic integrity and to quash specific threats.
“Our rules prohibit people from using Twitter for the purpose of manipulating or interfering in an election or other civic process,” Twitter spokesman Trenton Kennedy told CNN. “We won’t take action on broad, non-specific, or partisan statements about the integrity of elections. Our goal is to enforce our rules judiciously and impartially against clear attempts to undermine the election process.”
Trump’s baseless tweets about voter fraud were also posted onto Facebook, which hasn’t taken any action against any of his posts, which were subsequently shared by thousands.
Facebook founder Mark Zuckerberg has faced withering critiques for being too conciliatory toward Trump and not doing enough to crack down on misinformation and hate speech. He announced new policies in June that he said would more effectively “fight voter suppression” and provide “authoritative information on voting during the pandemic” among other things.
The strategy isn’t only about fact-checking or hiding problematic posts, which is a game of catch-up because Trump has more than 83 million followers and tweets at least a dozen times each day. Platforms can pro-actively combat misinformation by amplifying legitimate information from reputable sources and putting posts from election officials at the top of people’s feeds.
“The way to attack disinformation is to flood the zone with good information,” Persily said.
Twitter, Facebook and other sites took steps like these to deal with an wave of misinformation about the coronavirus. Platforms imposed new protections against posts about dangerous medical advice, and guided users toward trustworthy sources of information, like the Centers for Disease Control and Prevention.
For instance, when Trump suggested in April that injecting bleach into the body could cure Covid-19, Twitter quickly blocked the trending hashtags like #InjectDisinfectant.
Filling the federal void
Part of the problem is that the federal government has stepped back from its responsibilities.
The Trump administration has faced bipartisan criticism for not doing more to address foreign meddling in US elections. Congress hasn’t passed any comprehensive election security legislation in the wake of the 2016 debacle, which included a sweeping operation by the Kremlin to influence American public opinion through troll farms, fake accounts, and viral propaganda.
When it comes to bolstering election machines and equipment, Congress has approved only $400 million this year to use for states to make coronavirus-related changes. Voting rights experts believe the true price tag to properly prepare for November is closer to $2 billion.
Also, with the resignation of a top official, the Federal Election Commission has lost its quorum, rendering the watchdog powerless to police campaign finance laws. With campaign season now in full swing, the FEC isn’t currently able to take enforcement actions against rulebreakers
“If we had a functioning government, with nonpartisan overseers of elections, there would be an election authority who would play this role,” Persily said. “We don’t have that. Other countries do. So, we leave it up to these for-profit, multibillion-dollar corporations to make these calls.”
CNN’s Christian Pena and Donie O’Sullivan contributed to this story.