Inside Twitter’s decision to take action on Trump’s tweets

A weeks-old policy about virus misinformation laid a path for Twitter to push back on the president.

When Twitter Inc. earlier this month announced a new fact-checking tool, it was billed largely as a measure to combat false information about coronavirus. Two weeks later, the company deployed that tool in one of the biggest actions in its history: squaring off with President Trump.

Since Tuesday, the platform has taken several actions on messages from Mr. Trump as well as a post from the official White House account, marking some as breaking the company’s rules and adding a fact-check label to two about mail-in ballots. In response, Mr. Trump issued an executive order taking aim at what he said was censorship by social-media companies, and threatened to dismantle Twitter’s business if it didn’t stop tagging his posts.

The moves marked a sharp reversal for Twitter, which for years has faced criticism from users for what they see as inaction and inconsistency in policing its own platform, and turmoil among its employees over how it has managed its most prominent user. Some of Twitter’s 4,000 employees have accused the company and Chief Executive Jack Dorsey of ignoring harmful behaviour from the accounts of powerful figures, while others have said moderating those accounts would be akin to censorship.

Those tensions ratcheted up in recent weeks, say people familiar with the company’s operations, as the pandemic circled the globe and some of Mr. Trump’s frequent tweets to his 80 million followers tested Twitter’s rules.

Twitter censored Mr Trump’s now infamous “when the looting starts, the shooting starts” tweet.

“The internal decision-making wasn’t quite a domino effect, but it started slowly trickling,” an employee familiar with Twitter’s decisions said.

For years, employees across the company have met privately on an unofficial basis to discuss how the company could apply its rules to Mr. Trump’s postings, according to people familiar with the meetings. This past week’s decisions were unexpected, one of the people said.

Some Twitter staffers believed the company should enforce existing policies for reducing what it has categorised as hate speech and harmful content, even if involving Mr. Trump, who uses the service to comment on news and pop culture as well as make policy pronouncements.

Like Facebook Inc.’s Mark Zuckerberg, Mr. Dorsey has met privately with Mr. Trump and made efforts to cultivate relationships with prominent conservatives, but has also acknowledged that much of his workforce has left-leaning political views. In 2018, Mr. Dorsey said he didn’t feel Twitter’s conservative employees felt safe to express their opinions within the company.

The company’s head of site integrity, Yoel Roth, was cited by many conservatives this past week as an example of left-leaning bias at Twitter. In past tweets Mr. Roth has called Mr. Trump a “racist tangerine” and compared White House adviser Kellyanne Conway to Joseph Goebbels, the Nazi minister of propaganda. Mr. Roth was widely harassed and threatened online this past week after Ms. Conway called him out in a television interview.

In a tweet this past week, Mr. Dorsey said he accepted responsibility for the company’s decisions. “Fact check: there is someone ultimately accountable for our actions as a company, and that’s me,” he wrote, asking others not to attack Twitter employees in the wake of the moves against Mr. Trump.

Mr. Dorsey is an outspoken supporter of the Black Lives Matter movement but has long maintained that Twitter as a company is neither liberal or conservative. He has repeatedly told employees that people should be aware of the unfiltered views of public figures, current and former executives have said. That view angered large numbers of employees beginning in the campaign season of 2016, feelings that escalated after Mr. Trump won the White House and used Twitter to call adversaries names, mock their appearances and spread false information.

“You have to have rules you can enforce consistently, and that’s what they’ve always been lacking,” said one former employee who spoke with Mr. Dorsey directly about the matter.

Twitter has long said that maintaining a healthy public discourse on its platform is its priority. “Serving the public conversation includes providing the ability for anyone to talk about what matters to them,” the company has said previously in a statement. “This can be especially important when engaging with government officials and political figures.” For the past few years, Twitter employees said, their company appeared to be taking a relatively hands-off approach to moderating content. By contrast, Facebook had hired tens of thousands of content moderators and pledged to improve online conversation.

A role reversal of sorts began last year, when Twitter said it would ban political ads entirely, and Facebook made the decision to continue accepting them — and not to fact-check them. Mr. Zuckerberg talked more regularly about his commitment to free expression rather than stamping out toxic content.

More recent actions suggested Twitter was taking a different tack toward world leaders. In March, the platform deleted posts by Venezuelan President Nicolás Maduro and Brazilian President Jair Bolsonaro related to unproven methods of curing the new coronavirus.

And in April Twitter had added a fact-check label to a message posted by a Brazilian politician who claimed that quarantine increases the prevalence of COVID-19.

Still, many users and even some internally have felt the company’s enforcement of its rules has been inconsistent.

The fact-checking tool that Twitter unveiled in early May created a way for the company to flag the accounts of people in power without removing their content entirely. The company described the tool as a way to address conspiracy theories and other misinformation about the COVID-19 pandemic, but internally, employees noted that it could be used more broadly, the employee familiar with the company’s decisions said.

“We had never had a policy at Twitter that said you can’t lie or spread a conspiracy theory,” the employee said. “That was the starting point for this.” At first Twitter started marking links to content such as the conspiracy-laden Plandemic video as false. Then the company started thinking about how the policy could be applied to other critical topics, such as elections, the person said.

Twitter staff warned Mr. Trump’s team on May 20 that a tweet about voter fraud risked receiving a fact-check notice because it referred incorrectly to absentee ballots rather than absentee ballot applications. Mr. Trump deleted the tweet, The Wall Street Journal previously reported.

Then, on Tuesday, Twitter for the first time added fact-check labels to two tweets from Mr. Trump when he posted about the potential for fraud involving mail-in ballots.

Days later, on Friday, Twitter took an even bigger step when it added the label to Mr. Trump’s tweet that said the message promoted violence.

Mr. Trump had tweeted “when the looting starts, the shooting starts,” a line that some people took to refer to the former police chief of Miami when he cracked down on U.S. civil rights protests, as well as the former governor of Alabama, known for his opposition to the U.S. civil-rights movement.

Mr. Trump defended his message. “It was spoken as a fact, not as a statement,” he said in a subsequent tweet. “I didn’t want this to happen, and that’s what the expression put out last night means.” Twitter has a year-old policy of flagging tweets by political figures who violate its rules, although this was the first time the company had applied such a flag to a tweet from Mr. Trump.

“This tweet violated the Twitter rules about glorifying violence,” said the label affixed to Mr. Trump’s tweet. The label also said Twitter had determined it was in the public’s interest for the tweet to remain accessible, although the company restricted the way users could engage with the message.

The Wall Street Journal