‘We Make Mistakes’: Twitter’s Embrace of the Extreme Far Right

The following analysis lays out an introduction to observations the Southern Poverty Law Center (SPLC) has made about Twitter’s longstanding relationship with the far right. The analysis features samples of SPLC’s correspondence with Twitter about extreme far-right figures, as well as an insider’s revelations about the company’s struggles with moderation, and critical assessments of researchers from outside our organization. Tim Pool, a pro-Trump social media personality who has claimed to correspond with Twitter CEO Jack Dorsey in private, also told us when we reached out to him about the billionaire social media executive that he “seems very adamant that far-right figures be given unrestrained platforms.”

Twitter’s willingness to amplify extreme far-right voices has been a significant part of the company’s history. Now, those voices are one of the defining characteristics of the platform. The company created rules to prevent their users from spreading hateful content, employing automation or “bots” to boost tweets, and misleading the public about elections. Twitter does not enforce these rules with any discernible consistency. Dorsey and his staff have in fact enabled some repeat offenders, who post at a high volume on the site and have built up big followings to spread hate and disinformation. Many of these disinformation superspreaders have never faced any meaningful consequences for violating Twitter’s terms of service.

Hatewatch first reached out to Twitter for a comment on March 23 about the findings in our analysis, and the company said it would produce a statement. On March 24, Hatewatch agreed to a 30-minute, off-the-record conversation with three employees of the corporation to help them better understand what we decided to publish, at their request. On March 25, Twitter did not produce a statement, but instead contacted SPLC’s President and CEO Margaret Huang to protest the naming of employees in this analysis, citing security concerns. Hatewatch followed up with Twitter a second time, detailing additional reporting suggesting that its moderators “are powerless” and are beholden to decrees handed to them by what a source depicted to be the company’s ideologically driven leadership team. Twitter responded to that reporting by asking when the analysis would be published. Ultimately, the company chose not to provide a statement.

Twitter ‘is not going to do anything’

Hatewatch, the publishing arm of the SPLC’s Intelligence Project, has been in dialogue with Twitter about far-right accounts for years, often as it relates to reaching out for comment on stories. During that time, Twitter has suspended individual accounts and Hatewatch has repeatedly watched the personalities who operate them return to the site under new handles. Twitter has also shown little indication that it seeks to limit the proliferation of hate or disinformation on its platform in any systemic way. On the contrary, the Twitter business model appears to hinge on instilling feelings of resentment in people and, to at least some degree, exacerbating mental illness and anxiety. Extremists who terrorize other users and exploit the site to sow chaos keep the billion-dollar corporation’s business model humming.

“People have learned [that Twitter] is not going to do anything,” James Alefantis, the owner of Comet Ping Pong pizzeria in Washington, D.C., told Hatewatch in a phone conversation in February. Far-right personalities targeted Alefantis and his employees with harassment for more than four years, after Twitter enabled them to spread disinformation in the form of the infamous #Pizzagate conspiracy theory. Twitter has allowed some of the most infamous promoters of #Pizzagate to continue to push lies and hate on its platform years later, rewarding them with growing audiences. “The message ‘There’s nothing you can do [about Twitter]’ has been fully integrated into the American brain.”

Disempowering the moderators

Twitter CEO Jack Dorsey is sworn in as the Senate Select Committee on Intelligence prepares to hold an open hearing on Sept. 5, 2018, in the Dirkson Senate Office Building in Washington, D.C. (Photo by Pat Benic/UPI/Alamy News)

Twitter’s userbase has made CEO Jack Dorsey into one of the richest men on earth. He’s worth over $10 billion, thanks to his social media site, his financial services company Square and other projects. In addition to Twitter and Square, Dorsey also publicly supports the cryptocurrency Bitcoin, which he did in a series of tweets published on Jan. 13. That thread drifted through different topics, including Twitter’s decision to suspend former President Trump’s handle, @realDonaldTrump, following the Capitol violence his company helped enable.

Dorsey tweeted in January: “Yes, we all need to look critically at inconsistencies of our policy and enforcement. Yes, we need to look at how our service might incentivize distraction and harm. Yes, we need more transparency in our moderation operations.” But, Dorsey concluded in the same tweet thread, “All this can’t erode a free and open global internet.”

Dorsey’s vision of “a free and open global internet” appears to leave room for disinformation specialists to use Twitter’s traffic to destabilize democracy in the U.S. and attempt to push the country in a more illiberal direction. A source who is familiar with the inner workings of Twitter’s moderation told Hatewatch that Dorsey and some of the upper-level employees in his organization approach their work with an “activist, libertarian” ideology that drives them to allow people to push lies and hatred on the website. (Hatewatch is not naming its source to protect them from potential retaliation.)

The same source said many Twitter employees are “well-meaning,” and do not necessarily subscribe to the same libertarian ideology as the company’s leadership does. But the source also told Hatewatch that Twitter’s moderators “are powerless” and bound to decisions made by an ideologically driven leadership team. The moderators “feel pressure to follow [guidelines and instructions] to a T,” the source told Hatewatch. High-level decision-making about moderation policy comes from a different team, which tells moderators how to handle accounts. But the two teams are siloed, the source said, and rarely interact directly. Managers of Twitter’s moderators base the performance reviews of their staff on “how quickly they can clear their queues [of reported tweets and accounts] and how many [complaints] they can process in a day.”

An ‘activist, libertarian’ ideology

More ideologically driven managers at Twitter who decide who gets to operate on the platform, the source told Hatewatch, “view themselves as being the last line of protection from America from becoming China.” It’s the kind of absolutist worldview the fringe conspiracy theorists at Infowars sometimes voice.

Jack Dorsey himself “follows” or has followed on Twitter extreme far right and reactionary figures, as Hatewatch will detail later in this analysis. Though Dorsey follows nearly 5,000 accounts of varying political alignment, his connections to these far-right accounts show he cannot in good faith claim ignorance about the extremists who exploited their site in the runup up to the Jan. 6 attack. Not only did Twitter “verify” many of these personalities, but their CEO also publicly interacted with some of them. Beyond the people detailed here, Dorsey also once followed Stefan Molyneux, a social-media personality Twitter permanently suspended from the platform in 2020. Molyneux argued on Twitter that white men were genetically predisposed to be more intelligent than women and non-white people at the time Dorsey followed his account, Hatewatch found. (Molyneux has denied being a white supremacist, despite repeatedly remarking that non-white people are predisposed to be less intelligent than white people.)

Becca Lewis is a Ph.D. candidate at Stanford University who studies the way social media fosters and incentivizes far-right movements and enables them to spread propaganda. Lewis told Hatewatch that the source’s evaluation of Twitter’s apparent rightward ideological bent does not surprise her. Lewis noted to Hatewatch that major tech companies such as Twitter were founded on a libertarian ethos that today they disguise as being ideologically neutral.

“[Twitter] has an individualist view of speech,” Lewis said, referring to the belief that speech serves the will of the individual, rather than society. “It still deeply influences the way all of these social media companies do content moderation.”

Twitter hashtags

Twitter added #BlackLivesMatter and #BlackTransLivesMatter hashtags to its official account bio in summer 2020. (Screenshot via Twitter)

Twitter sometimes employs marketing tactics to project an image of progressivism, even though the website has developed a reputation for being littered with white nationalists. During the Black Lives Matter demonstrations in summer 2020, Twitter put #BlackLivesMatter and #BlackTransLivesMatter hashtags in its corporate bio on their site. The company removed the hashtags in September 2020, around the same time people started noticing the degree to which its algorithm cropped photos in such a way that it gave white faces a more prominent placement on its feed than Black ones. Twitter issued an apology over the photo cropping. Hatewatch tested the algorithm and found that it still favors white faces over Black ones.

Gamifying the discourse

The analytics company Similar Web lists Twitter as the sixth most trafficked website in the U.S., just behind Facebook, Amazon and Yahoo. Twitter, like all other major social media companies, designs its product to keep people tethered to it, hoping the screen time will result in people clicking. The company gamifies discourse to achieve that effect, targeting human psychology in much the same way casinos make choices in design to keep people gambling. Critics of Twitter, including SPLC, underscore the degree to which the traffic-hungry design produces the side effect of driving right-leaning users toward increasingly inflammatory content. A person may come to the site seeking to follow a conservative politician, then eventually be directed by Twitter’s algorithm to follow an account pushing white nationalist talking points.

“Even if the algorithm were removed, Twitter is filled with horrible content,” Megan Squire, a professor of computer science at Elon University and a Senior Fellow with Southern Poverty Law Center, told Hatewatch for this analysis. “Even if they removed the algorithm, the problem is what Twitter allows on its site. Inventing algorithms to promote that content is adding fuel on the fire.”

'Who to follow' suggestions on Twitter

In 2021 Twitter recommended in a “Who to follow” section the white nationalist talk show Red Ice, as well as one of its hosts, Henrik Palmgren, and white nationalist Scott Greer. (Screenshot via Twitter)

One example of how Twitter’s algorithm drags people to find what Squire calls “horrible content” is the degree to which the site regularly recommends users to follow white supremacists and other far-right extremists. Hatewatch found dozens of examples of Twitter recommending that people follow extremists, sometimes in a slate of three accounts at one time. In one example, Twitter recommended in a “Who to follow” section the white nationalist talk show Red Ice, as well as one of its hosts, Henrik Palmgren, and white nationalist Scott Greer. YouTube banned Red Ice TV in 2019, as did Facebook, and Palmgren and his cohost moved their talk show to a separate, independently operated video hosting site. By advertising this harder-to-find location on Twitter, Red Ice retained much of the mainstream visibility they should have lost at that time.

As for Greer, The Daily Caller fired him in 2018 in response to pressure over his ties to the white supremacist movement. According to archives, Twitter verified Greer’s account in 2015, when he still worked for The Daily Caller, and then did not remove his verification badge after the firing. In 2015, Greer had under 2,000 followers. He jammed the site with content, tweeting close to 45,000 times in under nine years, not including the possibility of deleted posts, and now boasts over 80,000 followers. Greer runs one of the many extant accounts on Twitter that promoted lies about the 2020 election, tweeting “fake news” after CNN projected Joe Biden as the winner.

Greer’s Twitter display name also “trended” on Jan. 13, a week after the insurrection, when he disparaged Rep. Alexandria Ocasio-Cortez, D-N.Y., along racial lines. The “Trending on Twitter” section displays the subjects or concepts that are being tweeted about in high volume on the site. People who provoke outrage on Twitter are commonly rewarded with increased visibility in the Trending section after site users react negatively to their comments.

Scott Greer retweet

Scott Greer referred disparagingly to Rep. Alexandria Ocasio-Cortez’s Puerto Rican heritage in this January tweet. (Screenshot via Twitter)

“The people with the strongest hate for Confederacy can’t trace any ties back to 1860s America,” Greer wrote of Ocasio-Cortez, referring to her Puerto Rican heritage.

An easily exploited game

Another way extremists exploit Twitter’s Trending section is to deliberately attach trending keywords to lies in order to gain unwarranted attention or to change people’s perceptions of an event as it is happening. Joan Donovan, the Research Director of Harvard’s Shorenstein Center on Media, Politics and Public Policy, told Hatewatch that one recent example of far-right Twitter posters manipulating the Trending section during a breaking news event happened on Jan. 6. Far-right posters used the feature to falsely suggest that antifa demonstrators attacked the Capitol, a lie that was repeated on the House floor later that day.

“When a breaking news event happens, there is this run towards whatever hashtag people [are coalescing around] in that moment. And it’s ripe for planting misinformation,” Donovan said in a video chat. “Nowhere was that more consequential than the Jan. 6 insurrection, because they were able to get that narrative to circulate very quickly that antifa was behind the Capitol [violence]. And now we are still reckoning with the consequences. Many people believe that it wasn’t Trump supporters at those gates.”

After the Jan. 6 attack on the Capitol, Twitter launched “Birdwatch,” a feature that tasks its own userbase with delivering Wikipedia-style annotations to misleading posts on an unpaid, volunteer basis. Birdwatch not only farms out labor to the human beings who produce Twitter’s content, but it also incentivizes them to remain logged in for longer hours in the hopes of cleaning up the site. Beyond numerous other limitations, including the fact that posters can simply block Birdwatch users from interacting with their tweets, the Trending on Twitter feature sometimes highlights disinformation immediately sitewide, giving volunteers little chance to blunt its spread at the same speed.

Jack Dorsey and Ali Alexander

Ali Alexander

Right-wing activist Ali Alexander takes a picture as he participates in former President Donald Trump’s “social media summit” with prominent conservative social media figures at the White House in Washington, D.C., July 11, 2019. (Photo via Reuters/Carlos Barria/Alamy News)

Ali Alexander, the far-right operator who galvanized the “Stop the Steal” protests after Trump’s electoral defeat, has for years touted a personal relationship with Dorsey. Alexander told The Epoch Times, an outlet that some have criticized for proliferating reactionary propaganda, in December 2020 that he “used [his] relationships with Twitter” to promote Stop the Steal events after the company initially took steps to limit him from doing so.

“Twitter originally banned our [Stop the Steal] link,” Alexander told The Epoch Times, referring to a website used to organize those anti-democratic protests. “I used my relationships with Twitter to get that reversed.”

Dorsey praised Alexander in a 2019 HuffPost story, saying that the notorious disinformation peddler made “interesting points.” Dorsey also acknowledged that he consulted with Alexander about whether to suspend the account of Infowars’ Alex Jones. (Twitter was the last major company to suspend Jones. They did it in September 2018, a full month after a cluster of companies, including YouTube, Facebook, Spotify and Apple, had already publicly cut ties with him.) Two years after Dorsey complimented Alexander’s point of view, the Stop the Steal leader used Twitter to call for revolution.

“If they do this, everyone can guess what me and 500,000 others will do to that [Capitol] building. 1776 is always an option,” Alexander’s tweet read. He was responding to a comment hate group-linked congressperson and Twitter celebrity Marjorie Taylor Greene made about politicians taking steps to ensure that Electoral College votes would be certified on Jan. 6. (Insurrectionists yelled “1776” while storming the Capitol building that day.) On the day of the insurrection, Alexander proclaimed that he did not disavow the violence that erupted.

Before Alexander led the political movement that eventually grew into the insurrection attempt on the U.S. Capitol, Dorsey and his team did not stop him from spreading disinformation. Ahead of the 2020 election, Alexander repeatedly promoted to his hundreds of thousands of Twitter followers a campaign called #JoeBidenIsSick, which proclaimed that Joe Biden, then the Democratic nominee for president, suffered from a degenerative illness he did not actually have. Dorsey and Twitter also permitted Alexander to promote a website that sold merchandise celebrating the lies about Biden’s health.

Ali Alexander tweet

Ali Alexander tweeted this message in April 2019. (Screenshot via Twitter)

In his last broadcast on Twitter’s livestreaming app, Periscope, Alexander threatened revenge against his perceived political opponents, saying he would “unleash a legion of angels to bring hell to our enemies.” Alexander made those comments on Jan. 10, at a time when government officials had boosted security across the country out of fears that extremists would hurt more people. Twitter finally removed him from the platform soon after.

Jack Dorsey and Mike Cernovich

Mike Cernovich

Mike Cernovich speaks at a Kelli Ward campaign event at the Gunsite academy in Ward’s bid to become the Republican nominee for an open Senate seat in Paulden, Arizona, Aug. 24, 2018. (Photo by Reuters/Conor Ralph/Alamy News)

Alexander is hardly the only Twitter-centric disinformation peddler involved in Stop the Steal protests, and he is also not the only one of whom Dorsey would be personally aware. Since at least 2017, Dorsey has followed the Twitter account for Mike Cernovich, an internet personality who has used the site to publish rape apologia and spread politically charged lies such as the #Pizzagate conspiracy theory. Beyond #Pizzagate, with its resounding lies about a pedophilia dungeon, Cernovich has weaponized Twitter to construct salacious frame-ups of liberals for years. Dorsey would likely be aware of this behavior. CBS News’s “60 Minutes” featured Cernovich on show that aired in March 2017 as an example of someone who manipulates Twitter to publish fake news.

Cernovich tweet

Mike Cernovich tweeted this message about #Pizzagate in November 2016. (Screenshot via Twitter)

“These weaponized social media attacks … they leave damage. Real, lasting damage in their wake,” Alefantis of Comet Ping Pong told Hatewatch about the lies promoted by Cernovich and others involved in #Pizzagate. “They brought a gunman to my restaurant. … My employees are traumatized, literally have PTSD, traumatized, waiters in their twenties. I know specific people who are going to therapists, or, you know, afraid to go to places because of the actions of these people. So, these are real consequences.”

Twitter added a “verification” badge to Cernovich’s account in summer 2016, according to archives, which is around the same time he allegedly organized a meeting of far-right extremists, including white nationalists, outside the Republican National Convention in Cleveland, as Hatewatch previously reported. Twitter first opened their verification to individual applicants in July of that year. Based on the general timing, Twitter verified Cernovich while he spread unsubstantiated conspiracy theories about Democratic nominee Hillary Clinton’s health.

“Hillary looks sick and unhealthy. She can barely talk. Her staff said she is ‘often confused.’ Only a rigged system would put her out ahead,” Cernovich tweeted in May 2016.

Days after the 2020 election, the California-based Cernovich showed up at a Stop the Steal event in Maricopa County, Arizona, among protesters that included armed militiamen. Cernovich branded the Stop the Steal event a “VOTER RIGHTS RALLY” on Twitter. Cernovich promoted the event to Twitter, including “pinning” a Stop the Steal livestream video to the top of his timeline. Since Dorsey is a follower, he could potentially have seen Cernovich spewing lies about the election in real time, just by opening his app and reading his feed.

“States that Trump won were not called early, but Arizona was. Even though left-wing consensus was that was a premature call. This was a psyop unlike any I’ve ever seen. This is stuff you [read] about and wonder, ‘How did it happen there.’ It’s happening here,” Cernovich tweeted, again playing into the fabricated narrative about election rigging.

Twitter did not respond to a request for comment about Dorsey’s interest in Cernovich’s account. Cernovich also did not respond to an emailed request for comment about this analysis. He describes himself on his Twitter bio as a “Stay at home dad of two daughters.”

Jack Dorsey and Tim Pool

Tim Pool

American journalist Tim Pool appears at a protest by anti-racist and anti-fascist groups against the far-right Democratic Football Lads Alliance’s “Day of Freedom” in London at which former English Defence League leader Tommy Robinson was scheduled to speak in May 2018. (Photo by incmonocle/Alamy Live News)

Dorsey also follows on Twitter a reactionary social media performer named Tim Pool, who uses his YouTube show to showcase far-right extremists such as Enrique Tarrio of the Proud Boys and the neo-Nazi collaborator Jack Posobiec. Election Integrity Partnership (EIP), a non-partisan group that includes researchers from Stanford University and the University of Washington, listed Pool among a group of verified Twitter “superspreaders” who pushed disinformation to Twitter following the 2020 election.

 

Pool tweet

Tim Pool referred to mainstream coverage of the 2020 presidential election in this tweet from December of that year. (Screenshot via Twitter)

“They kept saying there was no evidence of fraud,” Pool tweeted on Dec. 1, weeks after the election, apparently referring to the mainstream press. “Now they are moving on to ‘unproven conspiracies’ because there is evidence of fraud.”

Pool has also used his prodigious reach as a social media influencer to promote discussion about the idea that America is divided beyond repair and is doomed to descend into a second civil war. Pro-Trump insurrectionists wore T-shirts calling for a second civil war while entering the Capitol on Jan. 6.

Pool tweet

Tim Pool raised the specter of a new civil war in this August 2020 tweet. (Screenshot via Twitter)

“I’m feeling a bit bullish on Civil War. The realities have diverged to the point where you can’t bring them together,” Pool tweeted in August 2020.

Pool responded to a request for comment from Hatewatch by distancing himself from the far right and providing further context about his comments about the election and about the subject of “civil war.” Among other things, Pool stressed his mixed-race background and wrote that the “people who stormed the capitol deserve prison.” You can read Pool’s full reply by clicking here.

Dorsey joined Pool on “The Joe Rogan Experience” podcast in March 2019 to discuss Twitter’s moderation policies. Dorsey brought with him a woman named Vijay Gadde, the founder of a cryptocurrency investment firm who serves as his company’s Legal, Policy and Trust and Safety Lead. During the appearance, Gadde reassured Pool and Rogan that the company eventually sought to reinstate some far-right extremists they had previously suspended for violating their terms of service. Gadde called Gavin McInnes, the founder of the extreme far-right, street-brawling gang the Proud Boys, whom Twitter suspended after he used the site to celebrate political violence, a “test case for how we think about getting people back on the platform – in the long term.” Law enforcement has since arrested more than a dozen Proud Boys on charges related to the insurrection attempt on Jan. 6. McInnes is not among them.

On a Jan. 13 broadcast of his YouTube show, Pool claimed to have “chatted back-and-forth” with Dorsey using the site’s Direct Message service.

Twitter and Nick Fuentes

Nick Fuentes

Nick Fuentes appears at a “Stop the Steal” protest outside the Georgia Capitol on Nov. 18, 2020. (Photo by Nathan Posner/​Shutterstock)

Twitter verified the account of livestreamer Nick Fuentes in 2017, according to archives, after he attended the deadly “Unite the Right” rally in Charlottesville, Virginia, that August. At the time Twitter chose to verify Fuentes, he had no claim to fame outside of attending Unite the Right as a participant and running an independent talk show where he promoted a hard-right, authoritarian worldview.

Fuentes has since gained more than 135,000 followers on Twitter. He galvanized the Stop the Steal movement by inviting along his young fans to antidemocratic protests that he promoted on the platform. Extremists ultimately waved the “America First” flag associated with his livestream and movement inside the Capitol on Jan. 6. Fuentes himself was caught on camera outside the Capitol encouraging insurrectionists to “break down the barriers and disregard the police,” as Hatewatch previously reported.

Fuentes’ name recognition skyrocketed as a result of his participation in Stop the Steal. He uses Twitter to support his broadcasting career by redirecting viewers to new live-streaming websites through the site after he encounters roadblocks in the form of deplatforming. (Here are two examples of him using Twitter to hype content hosted on comparatively obscure websites: Example 1, Example 2.) Twitter is the main reason he has been successful in retaining his audience, according to his own telling.

“We retained 80{6557c92bab376e861f4db2362dd750ed9808ade9f2baf81ac39a444313a64dce} of the viewership,” Fuentes said on a Jan. 20 broadcast. “I mean that’s a pretty big deal. It’s hard to overstate what a big deal that is … just by saying, ‘Hey check me out on Twitter.’”

Fuentes frequently expresses fear that Twitter one day will suspend him, but the corporation’s employees told Hatewatch they have no plans to do so. Hatewatch emailed Twitter about Fuentes on Jan. 5, the day before the insurrection, alerting them to a clip from Fuentes’ show, which he livestreamed during impeachment proceedings against Trump, where he raised the specter of killing Republican lawmakers. “What can you and I do to a state legislator besides kill them?” Fuentes said. “We should not do that. I’m not advising that, but I mean what else can you do, right? Nothing.” Rather than address Fuentes’ account and his thinly veiled calls to violence or provide a substantive explanation about why they choose to welcome him, Twitter instead wrote back on Jan. 11, offering to enroll Hatewatch in their “Partner Support Portal.”

“At this point our enforcement team has not seen enough violative content from @NickJFuentes on Twitter to ban him. Hannah [Gais] and Michael [Edison Hayden]: has SPLC enrolled in our Partner Support Portal? If so, this is a great way to report violative content to us. If not, I’d be happy to get you enrolled,” Todd O’Boyle, a senior manager of public policy, wrote back to Hatewatch.

In a subsequent email, O’Boyle described the Partner Support Portal as a way for SPLC to volunteer help with their moderation.

“Our Partner Support Portal is an invite-only tool that allows partner organizations to report content they have directly to us. Upon enrolling @splcenter you would have access to the portal whenever you visited the Twitter Help Center while logged in as @splcenter. You can find additional details in this deck,” O’Boyle wrote in a follow-up email.

Twitter also provided at least two accounts to AFPAC, Fuentes’ yearly conference that runs alongside the more mainstream CPAC, although they have not verified them. Other major social media platforms, including YouTube and Facebook, have banned Fuentes for pushing hate in recent years, as have payment processing sites such as PayPal. Twitter is exceptional in this regard. DLive, a fringe livestreaming platform Fuentes used at the time he hyped Stop the Steal, suspended him during the fallout from Jan. 6, following two Hatewatch stories highlighting his use of their site. Clubhouse, an audio chat-based social media site, suspended Fuentes’ account in March, allegedly after he made comments mocking Black and disabled people, according to a reporter for The Daily Dot and Mediaite.

Fuentes has also used Twitter to falsely describe the Jan. 6 violence as a “false flag” event.

Fuentes tweet

Nick Fuentes described the Capitol insurrection as a “false flag” in this February tweet. (Screenshot via Twitter)

“Becoming more apparent as time has gone by that the Capitol Siege was just a big psyop/false flag,” he tweeted on Feb. 20.

Fuentes also tweets offensive or dehumanizing content and then deletes it, apparently to avoid being suspended.

“Aunt Jemima came off of the syrup bottle and onto the $20 bill,” he tweeted on Feb. 9, referring to the Black abolitionist Harriet Tubman.

Fuentes deleted the tweet after Twitter users denounced it as racist. Joan Donovan, the disinformation expert at Harvard’s Shorenstein Center, said this technique – posting inflammatory content, allowing it to provoke outrage and then deleting it before Twitter feels obliged to suspend the offending account – is a common tactic used by far-right actors. Hatewatch reached out to Fuentes by email for a comment on this investigation, but he did not respond. (Editor’s note: Following publication of this report, Twitter suspended Nick Fuentes’s account.)

Twitter and Jack Posobiec

Like Fuentes, Posobiec employs the tactic Donovan described, tweeting out inflammatory content and then deleting it to avoid being suspended. Hatewatch has written at length about Posobiec’s exploitation of Twitter. He operates arguably the most notorious account on the platform today. Twitter has protected Posobiec amid repeated violations of their terms of service, enabling the disinformation specialist to obtain more than 1 million followers.

Posobiec Twitter bio

Jack Posobiec claimed “FMR CBS News” in his bio in the months before Twitter verified his account. CBS News told Hatewatch that Posobiec never worked there. (Screenshot via Twitter)

Twitter verified Posobiec in April 2017, following six months of him claiming “fmr CBS News” in his bio. CBS News and two CBS affiliates stationed in cities where Posobiec lived told Hatewatch that he never worked for that company. Posobiec has used Twitter to target Jewish reporters with antisemitism, to hype a Polish neofascist group that in the 1930s bombed Jewish homes and to promote links to websites populated by neo-Nazis. Posobiec also spread the contents of a Russian intelligence-led hack across Twitter in the hopes of disrupting the outcome of the 2017 French elections and worked with neo-Nazis to make propaganda videos. Posobiec linked to an obscure junk news website called “SouthFront” 28 times in 10 months during the runup to the 2020 election, as Hatewatch previously reported. In April, the U.S. government labeled SouthFront as part of a Russian intelligence plot.

For years, Posobiec pushed the #StopTheSteal hashtag. He deleted those tweets after the violence on Jan. 6, but Hatewatch archived them. Posobiec’s followers also sometimes target people with harassment after he references them on the site in a negative context, as they did during #Pizzagate. In January, when a journalist reported on the Arizona Republican Party, whose Twitter account has also been accused of contributing to contributing to the atmosphere of right-wing violence that followed the 2020 election, Posobiec tweeted repeatedly about her personal life. After trolls swarmed her account, someone updated the reporter’s Wikipedia page to put “homewrecker” as her middle name. Her colleagues told Hatewatch they flagged Posobiec’s tweets, but Twitter took no action in response to the woman’s harassment. Hatewatch has chosen not to publish the name of the reporter to avoid triggering further harassment of her.

Posobiec tweet

Jack Posobiec promoted the “Stop the Steal” campaign concerning the 2020 presidential election in this September 2020 tweet, roughly two months before Election Day on Nov. 6. (Screenshot via Twitter)

Posobiec also involved himself in the on-the-ground Stop the Steal movement that later became the attack on the Capitol on Jan. 6. On the day of the violence itself, Posobiec made television content about it from a distance for the low-standard cable television station One America News Network. On Jan. 5 in Washington, D.C., Posobiec told The Epoch Times that the joint session of Congress the next day would be “something that no one’s seen before.”

“I think it’s going to be absolutely historic, and really codify America’s history not only for this time but for years to come,” he said, referring to what became an insurrection attempt.

Posobiec tweet

Jack Posobiec posted this picture in 2017 in the months before Twitter verified his account. (Screenshot via Twitter)

The last time Twitter responded to Hatewatch about Posobiec’s account, Twitter said that a fake story about a bomb threat that was retweeted over 30,000 times and a series of his tweets appearing to endorse vigilante violence against Black Lives Matter demonstrators made for acceptable content on the site.

“I’ve escalated the tweets you shared from Aug. 31 and they are not currently in violation,” Reggie McCrimmon of Twitter’s Global Public Policy team wrote to Hatewatch on behalf of the company.

McCrimmon wrote that Twitter would not respond to archives of tweets that Posobiec deleted. Internet archiving has been accepted as evidence in a court of law.

“Of the archived tweets you’ve previously shared can you please send their links on twitter and I will escalate for review. Once you share them and we review, I’ll schedule some time for us to chat,” McCrimmon wrote.

The source who detailed to Hatewatch the inner workings of Twitter’s moderation process said that people like McCrimmon are not given autonomy to make decisions about accounts and are instead beholden to decrees made by a siloed unit, with which he and others would have little communication. Twitter declined an opportunity to further clarify how they evaluate extreme far-right accounts like the one belonging to Posobiec. Hatewatch reached out to Posobiec by email for a comment on this investigation, but he did not respond to it.

Tweet labels do not limit exposure to bad information

Recently, Twitter has begun occasionally labeling viral posts that contain lies. Before suspending Trump’s account on Jan. 8, Twitter repeatedly put labels on posts of his that falsely claimed outside forces rigged the 2020 election to ensure his defeat. Extremists and disinformation peddlers commonly ridicule such labels, portraying them as empty and easy to ignore.

Twitter flagged a tweet by “verified” extreme far-right pundit Cassandra Fairbanks in that manner on Feb. 6. In Fairbanks’ original tweet, she hyped the conspiracy theory that mysterious ballots appeared to push Biden to victory and shared a link to a post she wrote about it on The Gateway Pundit, a junk news website. Twitter later labeled the Fairbanks post by noting that its claims of election fraud have been “disputed.” Fairbanks mocked Twitter’s label.

“Twitter suspended me for this and disabled likes and retweets to get people to stop looking at it… but it didn’t really work,” Fairbanks tweeted at about 2 a.m. Eastern time on Feb. 6.

Fairbanks’ apparent screenshot showed that Twitter users had viewed the flagged tweet more than 1,600,000 times in under 24 hours, demonstrating the impact and reach of the disinformation. Hatewatch reached out to Fairbanks for a comment on this investigation by email, but she did not respond. The propagandist uses Twitter at a staggering volume. Hatewatch calculated that during her first eight years on Twitter, Fairbanks tweeted around 50 times per day, including retweets and replies, and excluding possible deletions.

Fairbanks tweet

Cassandra Fairbanks tweeted this reference to the George Floyd protests in March. Among other things, she has used Twitter to read aloud from a terrorist manifesto. (Screenshot via Twitter)

“The George Floyd family should pay reparations to all the small business owners who had their shops attacked during the riots,” Fairbanks tweeted on March 12, referring to the grieving family of a Black Minneapolis man who died in police custody, following an incident in which an officer kneeled on his neck for nearly nine minutes.

Twitter verified Fairbanks as a public figure in the summer of 2016, around the time she became known on the site as a Bernie Sanders supporter who performatively switched to Donald Trump, and little else. Within a few months of receiving her blue check mark, Fairbanks also promoted the #Pizzagate lie on Twitter. Twitter suspended an account belonging to Jim Hoft in February. Hoft is the editor of The Gateway Pundit. Twitter chose to leave in place Fairbanks, arguably The Gateway Pundit’s most visible writer, who in addition to publishing disinformation used a Twitter livestream to read aloud from a terrorist manifesto, and also raised close to $25,000 for herself off a violent incident involving “antifa” that appears not to have happened, according to reporting from Right Wing Watch. In April, she tweeted that convicted murderer Derek Chauvin “is a political prisoner,” echoing white supremacist talking points about that case. Fairbanks has amassed more than 250,000 Twitter followers since joining the site in 2012.

Twitter suspensions mean little to extremists

Twitter suspensions typically represent only a temporary pause for extremists who use the site to organize or promote propaganda. When discussing this topic with Hatewatch, Brian Friedberg, a senior researcher also with Harvard’s Shorenstein Center, joked with Hatewatch about “Mel Gibson Fan,” a pseudonymous antisemitic and racist troll who tweets hate from an account for weeks at a time before Twitter suspends them and they switch handles. Each time Twitter suspends an iteration of Mel Gibson Fan, they return with a new number next to their display name: Example 1, Example 2, Example 3. As of February, Mel Gibson Fan appeared to their followers as “Mel Gibson Fan 66,” helping researchers keep track of how many times he’s been suspended and then returned to the site.

“Pseudonymous accounts and ban evasion have long been tried and true tactics employed by dedicated anti-Black, antisemitic and misogynistic harassment networks on Twitter. These factions easily manipulated Twitter’s functionality to remain online and gamified the spread of racism and far right ideology on the platform,” Friedberg told Hatewatch by email.

Spectre tweet

Twitter user @TheSpectreOps posted this response to the 2018 shooting at the Capital Gazette newspaper in Annapolis, Maryland, that killed five. The same user has returned to Twitter dozens of times after being suspended. (Screenshot via Twitter)

Independent social media researcher Erin Gallagher published a Medium post in February on members of “The Shed,” a notorious collection of Twitter harassers who return repeatedly to the site under different handles. One of them, who goes by the pseudonym “Spicci,” has returned to Twitter at least scores of times in order to harass people, as Gallagher noted. (The same user bragged to Hatewatch about evading Twitter bans hundreds of times in a private conversation held over direct message in 2018.) Spectre, a white supremacist troll Hatewatch identified as Norman Asa Garrison III of Texas in a 2019 story, has similarly churned through numerous handles on Twitter, seeking to continue his abuse of Dorsey’s other users.

In the run-up to the 2020 election, Garrison also used Twitter to promote a reactionary, self-run website called The Lancaster Patriot, which focused on the swing state of Pennsylvania. A notorious pro-Kremlin propagandist named Charles Bausman helped support that site, as Hatewatch previously reported. Twitter removed an account associated with The Lancaster Patriot after Hatewatch reached out for comment for that Oct. 7, 2020, story but not before verified, high-follower accounts boosted Garrison’s hyper-partisan content and Fox News credited it during a television broadcast. Bausman himself has maintained multiple Twitter handles over the last few years.

Affiliates of Mike Peinovich’s white nationalist group The Right Stuff, including Garrison, have for years burned through Twitter handles while promoting Holocaust denial, racism, anti-LGBTQ bigotry and fascism. Another prolific Twitter user from Peinovich’s orbit is a neo-Nazi named Joseph Jordan, who edits and writes for a junk news website. Jordan has repeatedly returned to Twitter under different handles related to his stage name, “Eric Striker.” Suspended State Department official Matthew Q. Gebert, a white nationalist also linked to Peinovich, has returned to Twitter under roughly a dozen different handles related to his racist worldview. Most recently, he returned to the platform to promote a white genocide-themed podcast.

“This week we’re honored to welcome Patriot Front founder Thomas Rousseau to break down the group’s ethos and objectives,” Gebert tweeted on Dec. 4, 2020. Patriot Front is a white supremacist hate group, and Rousseau used his appearance on the podcast, which Gebert “pinned” to his Twitter profile, to solicit new members.

On Dec. 3, 2020, Hatewatch flagged those recent accounts operated by Jordan and Gebert to Twitter’s Public Policy department. Twitter left the accounts up until at least mid-February. Archives show that between the time Twitter became aware of Jordan posting to the site and when they suspended that account, he tweeted hundreds of times per month, including retweets. In Gebert’s case, the posts exceeded a rate of over 1,000 times per month.

‘We make mistakes’

When Dorsey, along with Facebook’s Mark Zuckerberg and Google’s Sindar Pichai testified before Congress in a virtual hearing on March 25, the Twitter executive appeared to acknowledge that his company deserves at least some blame for the violence on Jan. 6. Dorsey was also the only one of the three executives to answer in the affirmative to a question about whether his company helped spread disinformation that led to the Capitol attack.

“We are a bunch of humans with a desire to make the world around us better for everyone living today and for those that came after us. We make mistakes in prioritization and in execution,” Dorsey said in his opening statement.

What Dorsey ignored is that the “mistakes” on Twitter that led to the Capitol attack preceded the insurrection by over four years. Long before it became a political movement capable of unleashing violence, high-profile Twitter posters made #StopTheSteal a popular hashtag on the site, applying it to a range of elections as a way of eroding faith in the democratic process, starting with the 2016 presidential election and extending to the 2018 midterms. Hatewatch calculated that after the 2020 election, following the lead of Twitter’s army of right-wing influencers, pro-Trump Twitter users pushed #StopTheSteal dozens of times per minute. (You can read some background about how #StopTheSteal was used on Twitter by clicking here.)

The Atlantic Council’s DFRLab published to the website JustSecurity a timeline of how Stop the Steal evolved into the insurrection, also noting that the movement started specifically on Twitter. Posobiec posted to the site about #StopTheSteal as it related to the 2020 election as early as Sept. 7, appearing to kick off what would become the attempt to overturn its eventual results, the timeline shows.

Jared Holt, a resident fellow at DFRLab who worked on the JustSecurity timeline, told Hatewatch: “Although some of the most inflammatory and violent rhetoric we observed prior to the attack happened in other online venues, the participants of those conversations were often reacting to conspiracy theories and disinformation first shared to Twitter. It’s hard to imagine that Jan. 6 would have been what it was without the site.”

Roger Stone, Posobiec’s mentor, who has appeared at events alongside members of violent far-right groups that participated in the attack such as the Proud Boys and the Oath Keepers, first popularized the hashtag #StopTheSteal in 2016. He continued to use it on Twitter in runup to the attack on the Capitol on Jan. 6 without encountering any hindrance. Twitter suspended Stone’s verified account in October 2017, but Dorsey’s company permitted him to continue to access the website through an account advertising a podcast called “Perfectly Clear w/ Roger Stone.”

Stone tweet

An account associated with Roger Stone’s podcast tweeted this endorsement of the #StopTheSteal campaign just before the Jan. 6 insurrection at the U.S. Capitol. (Screenshot via Twitter)

“SUPPORT the #STOPTHESTEAL SECURITY PROJECT at stopthesteal.org,” Stone’s podcast account tweeted on Dec. 30, promoting an organizing link to the event that led to the Capitol violence.

Other high volume, large follower-count Twitter accounts belonging to such right-wing activists as Scott Presler and Brandon Straka also pushed #StopTheSteal. Like Posobiec, Twitter awarded both of those men verification badges despite their lack of discernible credentials outside of activism such as #StopTheSteal. The FBI arrested Straka on Jan. 20 on charges related to the insurrection. In his indictment, the word “Twitter” appeared 13 different times.

“[Be] embarrassed & hide if you need to- but I was there. It was not Antifa at the Capitol. It was freedom loving Patriots who were DESPERATE to fight for the final hope of our Republic because literally nobody cares about them. Everyone else can denounce them. I will not,” Straka tweeted on Jan. 6, according to the indictment. Twitter never took action against Straka’s account, even after his arrest.

As the March 25 hearing went on, observers accused Dorsey of abandoning his contrite tone for one of flippancy. He at one point diverted his attentions from the proceedings to tweet out a “Yes/No” poll, which critics interpreted as a “troll” of a Congressperson who pressed the social media executives to respond in a straightforward, yes-or-no manner. Then, a day after the hearing ended, Cernovich, Dorsey’s mutual follow, who pushed #Pizzagate and also hyped Stop the Steal events, said more about Twitter in one line than a critic of the website ever could.

“Twitter is an unqualified net social positive,” Cernovich, who once used the platform to remark that Trayvon Martin “got got before he was able to rape anyone,” wrote of his preferred platform.

Photo illustration by SPLC