By Keach Hagey and Jeff Horwitz 

Parler launched in 2018 as a freewheeling social-media site for users fed up with the rules on Facebook and Twitter, and it quickly won fans from supporters of President Trump. On Monday, it went dark, felled by blowback over its more permissive approach.

Amazon.com Inc. abruptly ended web-hosting services to the company, effectively halting its operations, prompting Parler to sue Amazon in Seattle federal court. Other tech partners also acted, crippling operations.

Driving the decision was the role of Parler in last week's mob attack on the U.S. Capitol. On the afternoon of the riot, Amazon warned executives from Parler it had received reports the social-media platform was hosting "inappropriate" content, and that Parler had 24 hours to address it.

"We have been appropriately addressing this type of content and actively working with law enforcement for weeks now," Parler policy chief Amy Peikoff told Amazon a few hours later in an email reviewed by The Wall Street Journal.

Amazon wrote back Thursday: "Please consider it resolved." The email gave Parler executives confidence that their moderation system, however strained, was acceptable to its tech partner.

It wasn't. Within two days of that correspondence, Amazon announced it was booting Parler from its cloud platform, joining Alphabet Inc.'s Google and Apple Inc. in pulling the plug on the service. Other vendors turned their backs, too: Twilio Inc. cut off Parler's two-factor authentication system, preventing it from weeding out fake new accounts, and Okta Inc. locked Parler out of key enterprise software tools.

Parler had become a pariah for serving as a hub for people who organized, participated in or celebrated the storming of the Capitol that left five dead, as well as a forum for some who have posted about future violent actions around the coming inauguration.

In some Parler posts flagged by tech companies as examples of inadequate moderation, users posted about "poisoning the water supply" in minority neighborhoods and killing their alleged enemies.

Executives at Parler said posts inciting violence violate its rules, although acknowledged they are aware such content was on the platform and others in the run-up to the Capitol attack. Its team of moderators -- mainly volunteers who receive training on what content should be removed -- has been overwhelmed and often faced large backlogs in handling offending posts as the service added users by the thousands in recent months, they said.

Parler Chief Executive John Matze said the company had offered to use algorithms to help moderators identify and weed out violent content, tools employed by larger social networks that Parler executives have previously resisted.

Parler is in a frenzied push to find new vendors to host its services so that it can resume operations, a process executives said could last at least a week. Backed by investors such as Republican donor Rebekah Mercer, Parler has the resources to restore service outside of Amazon and plenty of cash on hand, said Mr. Matze. But he acknowledged that a sustained loss of service could undermine the platform's future.

"We had regular conversations with them, and none of them gave us any indications that there were deep problems in the relationship," said Jeffrey Wernick, Parler's chief operating officer.

The company's looser policies on content moderation have attracted large groups of supporters of President Trump who use it to push their claims that the 2020 election was stolen, particularly as larger platforms like Twitter and Facebook have cracked down on this kind of speech. The company says it has 15 million users.

Social-media companies of all stripes have struggled to strike the right balance between free speech and content moderation. Liberals have generally argued that the tech platforms should be more aggressive in policing hate speech, while conservatives have complained that Big Tech is biased against their point of view.

Parler's troubles show the power a small number of tech companies have over online discourse. Moves made by platforms including Twitter and Facebook to ban President Trump, citing rules prohibiting content that incites violence, have the potential to reshape their businesses.

Before Parler was shut down, users alerted followers to reach them on other platforms, including those on Gab, or on messaging apps such as Telegram. Monday morning, Parler's website, parler.com, was down and users could no longer access news feeds or make new posts.

Hours later, Parler filed its lawsuit, alleging it was kicked off Amazon's cloud servers due to "political animus" and for anticompetitive reasons. An Amazon spokesman said the lawsuit's claims had no merit and that it respects Parler's right to determine what content it will allow.

Amazon explained its decision to shut off services to Parler by citing 98 instances of violent content on the platform that it said violated its rules. Parler said it removed all 98 items, in some instances before Amazon reported them.

Determining whether discourse on Parler or the platform's moderation of it was fundamentally worse than on other forums is "kind of an impossible question, empirically and philosophically," said Evelyn Douek, a Harvard Law School lecturer who studies content moderation. While a case can be made for app stores and other internet infrastructure providers taking action against platforms that don't take moderation requirements seriously, Ms. Douek said, the speed of tech companies' action against Parler raised questions.

"If 98 is the threshold, has AWS looked at the rest of the internet," she said. "Or at Amazon?"

Daniel J. Jones, president of nonprofit research group Advance Democracy, said the organization found numerous Parler users calling for violence on Jan. 6.

"Far-right extremists and conspiratorial groups, such as QAnon, specifically flocked to Parler because of the lack of moderation and guidelines," he said.

Parler relies on a volunteer community of jurors to moderate content. After some experimentation with its initial rules, Parler settled on generally allowing users to engage in any constitutionally protected speech. There would be no fact-checking, no restrictions on offensive language and no prohibition on gory or adult content, so long as it was tagged "sensitive" by the creator, executives said. Threats, spam and criminal activity weren't allowed, they said.

The first big test of the system came in November, when conservatives upset about the outcome of the presidential election flocked to Parler, doubling the company's user base to more than 10 million within a few days. The new users brought a host of problems such as spam and pornography, executives said, prompting complaints from larger tech platforms. The volunteer jury system seemed to work, they said.

To keep up, Parler tripled its volunteer moderator staff from 200 to 600 over the past two months and started paying them monthly stipends. It also has been hiring full-time moderators in recent weeks.

The Journal viewed some comments flagged for review under Parler's moderation system. The content included spam ads for Trump commemorative coins, fake accounts and incitement to violence.

"Threatens to kill Pence," one user noted about a post. Another, flagging a call to kill Facebook CEO Mark Zuckerberg and Twitter CEO Jack Dorsey, said: "You can't allow this or we won't have a voice."

Mr. Matze and other Parler executives noted that content calling for the overthrow of the U.S. government remains on other platforms. The Wall Street Journal has previously written about a private Facebook group called "the 2020 Civil War," where users discussed plans to march on the White House on Inauguration Day.

Facebook said on Monday it was removing content from its flagship platform and Instagram containing the phrase "stop the steal" -- a reference to claims that Mr. Trump would have won the 2020 election if not for widespread election fraud.

On Saturday, "HangMikePence" was trending on Twitter, which reflected chants from protesters at the Capitol who were disappointed with the vice president's decision to proceed with certifying election results. Twitter said Saturday evening that it blocked the phrase and other similar ones from trending to promote healthy discussions on the platform.

On Dec. 14, Amazon flagged four posts to Parler, saying the content "clearly encourages or incites people to commit violence against others, " which was a violation of its terms of service, according to an email reviewed by the Journal. One of the posts calling for violence was from Nov. 16. Another, from early December, included comments such as: "My wishes for a racewar have never been higher. I find myself thinking about killing n -- s and jews more and more often."

Ms. Peikoff said Parler removed the content immediately and believed Amazon was satisfied. One person familiar with Amazon's thinking said that correspondence began a process of enforcement that ended with Amazon cutting ties to the social network.

Parler executives said before the Capitol riot there was a growing number of calls for violence on the platform, as well on the rest of the internet.

Ms. Peikoff said she instructed moderators who had been hunting for spam to also look for incitements of violence and report it to law enforcement when appropriate. "I was concerned that there was actually going to be some sort of violence on the 6th," she said.

The day of the riot, Parler was among the social-media networks, including Facebook, YouTube and Twitter, that Trump supporters seeking to contest the election results used to organize the protests and celebrate the attack, according to researchers and analysts who study extremism and disinformation, as well as posts reviewed by the Journal.

Unlike those services, Parler had actively courted users disenchanted with the major platforms. Mr. Matze acknowledges that Parler's moderators fell behind, given the surge of activity last week Wednesday, by around 20,000 reports -- a roughly two-day backlog.

By the next day, advocates such as Sleeping Giants, a Twitter account created to press advertisers to stop supporting what they consider hate speech, and tech companies' employees were putting pressure on Google, Apple, Amazon and others to sever ties with Parler.

In suspending Parler from the Play Store, Google said the company needed to implement a robust moderation system to handle "egregious" content.

Apple told Parler it received numerous complaints that the platform had been used to organize the assault on the Capitol and was being used to organize future violence. Its list of evidence began with a link to Sleeping Giants' feed, which had screenshots of posts from influencers such as pro-Trump lawyer L. Lin Wood calling for firing squads to shoot Mr. Pence to an account with few followers calling for people to bring their weapons to the nation's capital on Jan 19.

Parler executives said in interviews Saturday that all the posts flagged by Apple and most of these accounts had been taken down. Mr. Wood told the Journal Sunday his post was meant as rhetorical hyperbole.

--Georgia Wells contributed to this article.

Write to Keach Hagey at keach.hagey@wsj.com and Jeff Horwitz at Jeff.Horwitz@wsj.com

 

(END) Dow Jones Newswires

January 11, 2021 20:43 ET (01:43 GMT)

Copyright (c) 2021 Dow Jones & Company, Inc.