BREAKING: Trump was censored by Twitter ‘under pressure from federal agencies’ before being banned

On Friday, journalist Matt Taibbi released the third installment of the Twitter Files, this time pertaining to the banning of former President Donald Trump from the social media platform shortly after the riots at the Capitol building on January 6, 2021.

Taibbi said that there was an "erosion of standards" within Twitter in the months leading up to the events of January 6, 2021, "decisions by high-ranking executives to violate their own policies," and more, going on "against the backdrop of ongoing, documented interaction with federal agencies."

Friday’s drop, Taibbi said, covers a span of time reaching from before the 2020 presidential election through January 6.

"Whatever your opinion on the decision to remove Trump that day, the internal communications at Twitter between January 6th-January 8th have clear historical import. Even Twitter’s employees understood in the moment it was a landmark moment in the annals of speech," wrote Taibbi.

A screenshot of an internal communication stated: "Is this the first sitting head of state to ever be suspended?"

"As soon as they finished banning Trump, Twitter execs started processing new power. They prepared to ban future presidents and White Houses – perhaps even Joe Biden. The 'new administration,' says one exec, 'will not be suspended by Twitter unless absolutely necessary,'" Taibbi wrote.

An internal communication, citing Twitter’s rules on ban evasion, stated "if it is clear that another account is being used for the purposes of evading a ban, it is also subject to suspension. For government accounts, such as @POTUS and @WhiteHouse, we will not suspend those accounts but will take action to limit their use."

"However, these accounts will be transitioned over to the new administration in due course and will not be suspended by Twitter unless absolutely necessary to alleviate real-world harm," the message added.

Taibbi said that Twitter executives had removed Trump "in part over what one executive called the 'context surrounding': actions by Trump and supporters 'over the course of the election and frankly last 4+ years.' In the end, they looked at a broad picture. But that approach can cut both ways.

An internal message to former head of legal, policy, and trust at Twitter Vijaya Gadde, stated, "Hi Vijaya - I’m working with [redacted] on my team to put together a doc to share with you with a POV from research (ours, academics with whom we have been working, etc.) on DJT’s language as coded incitement to further violence."

A follow-up message stated, "In the mean time, here is our quick take: the decision on whether to pull that particular tweet or use that as a last straw for trump depends on many factors including: (1) the overall context and narrative in which that tweet lives."

"We currently analyze tweets and consider them at a tweet-by-tweet basis which does not appropriately take into account the context surrounding," the message continued. "You can use the yelling fire into a crowded theater example - context matters and the narrative that trump and his friends have pursued over the course of this election and frankly last 4+ years must be taken into account when interpreting and analyzing that tweet."

The message added a second point: "the larger question is around our moral imperative and decision as a company, which user sentiment should not drive based on #1 [redacted] and I believe that his tweet does violate our rules when taking that historical context + current climate into account."

Taibbi said that most debate surrounding Trump’s ban "took place in those three January days."

"However, the intellectual framework was laid in the months preceding the Capitol riots," he added.

Noting Twitter before January 6, Taibbi said that Twitter was "a unique mix of automated, rules-based enforcement, and more subjective moderation by senior executives."

"As the election approached, senior executives – perhaps under pressure from federal agencies, with whom they met more as time progressed – increasingly struggled with rules, and began to speak of 'vios' as pretexts to do what they’d likely have done anyway," Taibbi wrote.

Following January 6, internal messages within Slack showed Twitter executives "getting a kick out of intensified relationships with federal agencies," Taibbi said.

In these messages, between Yoel Roth and another person whose name has been redacted, Roth said, "Ehh, it happens. I’m a big believer in calendar transparency. But I reached a certain point where my meeting have become… very interesting… to people and there weren’t meeting names generic enough to cover."

"Very Boring Business Meeting That Is Definitely Not About Trump," a person responded.

"Preeeeeeeetty much," Roth responded, adding, "DEFINITELY NOT meeting with the FBI I SWEAR."

"These initial reports are based on searches for docs linked to prominent executives, whose names are already public. They include Roth, former trust and policy chief Vijaya Gadde, and recently plank-walked Deputy General Counsel (and former top FBI lawyer) Jim Baker," wrote Taibbi.

Taibbi noted a particular Slack channel which he said "offers an unique window into the evolving thinking of top officials in late 2020 and early 2021."

This channel, called "us2020_xfn_enforcement," was opened on October 8, 2020. Through January 6, this channel "would be home for discussions about election-related removals, especially ones that involved 'high-profile’ accounts (often called 'VITs' or 'Very Important Tweeters')," said Taibbi.

An introductory message to the channel read: "Hey Everyone… starting tomorrow (October 9th) until November 15th this channel will be used for the following reasons related to the US 2020 Elections."

Some of these reasons were "Trends Identified that require scaled investigations" and "High Profile Account Escalation that potentially require PII/Soft Intervention."

Included in this channel were Site Integrity, Safety Policy, Product Trust, Safety Operations, Media Ops, and  Global Escalation Teams.

"There was at least some tension between Safety Operations – a larger department whose staffers used a more rules-based process for addressing issues like porn, scams, and threats – and a smaller, more powerful cadre of senior policy execs like Roth and Gadde," wrote Taibbi.

Taibbi described the latter group as a "high-speed Supreme Court of moderation, issuing content rulings on the fly, often in minutes and based on guesses, gut calls, even Google searches, even in cases involving the President."

In an internal message between a person whose identity has been redacted and Roth, they spoke of a tweet from Trump on October 9, 2020 that stated: "Breaking News: 50,000 OHIO VOTERS getting WRONG ABSENTEE BALLOTS. Out of control. A Rigged Election!!!"

"'A rigged election would be enough to be run violation right?" The unidentified person wrote.

"If the claim of fact were inaccurate, yes," responded Roth, who added, "But it looks like that’s true," noting a news story covering what Trump’s tweet was about.

Taibbi stated that during this time, executives with the company were "clearly liaising with federal enforcement and intelligence agencies about moderation of election-related content."

One person asked Policy Director Nick Pickles in an internal message if he is "comfortable with Marketing talking about misinformation by saying that we detect it through ML, human review and **partnerships with outside experts?*"

"I know that’s been a slippery process, so not sure if you want our public explanation to hang our hat on that," they added.

In response, Pickles said, "can we just say 'partnerships'" and "eg not sure we’d describe the FBI/DHS as experts, or some NGOs that aren’t academic."

One post regarding the Hunter Biden laptop story revealed that Roth met weekly with both the FBI and DHS as well as the Office of the Director of National Intelligence.

"Hacked Materials exploded. We blocked the NYP story, then we unblocked it (but said the opposite), then said we unblocked it… and now we’re in a messy situation where our policy is in shambles, comms is angry, reporters think we’re idiots, and we’re refactoring an exceedingly complex policy 18 days out from the election."

The message revealed a "weekly sync with FBI/DHS/DNI" regarding election security.

"Some of Roth’s later Slacks indicate his weekly confabs with federal law enforcement involved separate meetings. Here, he ghosts the FBI and DHS, respectively, to go first to an 'Aspen Institute thing,' then take a call with Apple," wrote Taibbi.

One message from Roth revealed that he had to miss an FBI and DHS meeting, stating that he has "an Aspen Institute thing this morning on vaccines that I have to present at, and then a call with Apple to avoid us getting kicked out of the App Store during the DHS one."

In one case, the FBI sent a pair of tweets, one regarding a former Tippecanoe County, Indiana Councilor and a Republican named John Basham with claimed "Between 2% and 25% of Ballots by Mail are Being Rejected for Errors." The FBI cited a Politifact fact-check, saying that the claim had been "proven false."

The tweet was circulated in the enforcement Slack, with the group deciding to apply a "Learn how voting is safe and secure" label to it.

One commenter said, "it’s totally normal to have a 2% error rate," to which Roth gave the approval.

Taibbi noted that in examining the enforcement Slack channel, "we didn’t see one reference to moderation requests from the Trump campaign, the Trump White House, or Republicans generally. We looked. They may exist: we were told they do. However, they were absent here."

Regarding a tweet from former Arizona Governor Mike Huckabee in which he joked about mailing in ballots for his "deceased parents and grandparents," it sparked a long Slack in which Roth stated, "I agree it’s a joke, but he’s also literally admitting in a tweet a crime."

The group declared that Huckabee’s case was an "edge case," and though one employee noted "we don’t make exceptions for jokes or satire," they left Huckabee alone because "we’ve poked enough bears."

One person noted that Huckabee’s post could "still mislead people," to which Roth responded that regulation could depend on whether the joke would result in "confusion."

"In the docs, execs often expand criteria to subjective issues like intent (yes, a video is authentic, but why was it shown?), orientation (was a banned tweet shown to condemn, or support?), or reception (did a joke cause 'confusion’?). This reflex will become key in J6," Taibbi wrote.

Another example came in the form of a tweet from Trump, in which employees were prepared to put a similar "mail-in voting is safe" label on it. This tweet was about a mail issue in Ohio, and employees eventually realized that the "events took place," which meant that the post was "factually accurate." 

Taibbi revealed that Trump had been "visibly filtered" as late as one week before the election. "Here, senior execs didn’t appear to have a particular violation, but still worked fast to make sure a fairly anodyne Trump tweet couldn’t be 'replied to, shared, or liked,’" Taibbi wrote.

"Very well done on speed folks, what this is all designed for and a huge positive for the platform," one employee responded.

Actor James Woods tweeted a screenshot of the warning label on one of Trump’s posts, to which Twitter employees "despaired of a reason for action, but resolved to "hit him hard on future vio,’" wrote Taibbi.

One person wrote that they "action" Woods over "something worth the fiasco rather than this screenshot, since we don’t have a firm policy basis for action on his account."

In regards to a post made by Georgia Republican Congresswoman Jody Nice, Twitter teams suggested only "soft intervention," with Roth being concerned about "wah wah censorship."

Taibbi noted that numerous Biden supporting tweets warning that Trump "may try to steal the election" made their way to the team, only to have them refuse to take action on them.

Even in regards to a hashtag called "StealOurVotes," the team allowed the posts to remain, stating that it’s "understandable, and a "reference to a US Supreme Court decision on processing mail-in ballots that arrive after Election Day."

On December 10th, 2020, twitter executives announced a new "L3 deamplification" tool which now meant that a warning label applied to a tweet would be accompanied by deamplification. This came in the midst of a tweeting spree by Trump.

"Some executives wanted to use the new deamplification tool to silently limit Trump’s reach more right away," wrote Taibbi, beginning with a December 10, 2020 tweet from Newsmax regarding an interview with Rep. Jim Jordan.

"The significance is that it shows that Twitter, in 2020 at least, was deploying a vast range of visible and invisible tools to rein in Trump’s engagement, long before J6. The ban will come after other avenues are exhausted," wrote Taibbi.

Taibbi continued on to explain that in these files, Twitter executives frequently speak of "let’s put a bot on that."

"A bot is just any automated heuristic moderation rule. It can be anything: every time a person in Brazil uses 'green’ and 'blob' in the same sentence, action might be taken," he explained.

In one case, moderators added a bot to a Breitbart video tweeted by Trump in which the former President spoke on election fraud.

"The bot ends up becoming an automated tool invisibly watching both Trump and, apparently, Breitbart ('will add media ID to bot’). Trump by J6 was quickly covered in bots," wrote Taibbi.

"There is no way to follow the frenzied exchanges among Twitter personnel from between January 6thand 8th without knowing the basics of the company’s vast lexicon of acronyms and Orwellian unwords," wrote Taibbi.

To "bounce" an account, Taibbi explained, is to put an account in a timeout, usually spanning 12 hours.

"Interstitial" means "placing a physical label atop a tweet, so it can’t be seen," wrote Taibbi.

PII could mean "Public Interest Interstitial," meaning a covering label applied for "public interest reasons."

"This is all necessary background to J6. Before the riots, the company was engaged in an inherently insane/impossible project, trying to create an ever-expanding, ostensibly rational set of rules to regulate every conceivable speech situation that might arise between humans," wrote Taibbi.

"The firm’s executives on day 1 of the January 6th crisis at least tried to pay lip service to its dizzying array of rules. By day 2, they began wavering. By day 3, a million rules were reduced to one: what we say, goes," he added.

As the events on January 6 began to unfold, Taibbi said that there were "frantic calls’ for Twitter to start using its "full arsenal of moderation tools."

"What is the right remediation? Do we interstitial the video?" One employee asked.

At 2:39 pm PST, a communications official based Roth to confirm or deny whether the company had restricted Trump’s ability to tweet.

Roth responds: "we have not."

Just minutes later, Roth bounced Trump. An employee responded, "It’s the right decision. Yoel, I hope you, Del, and others are appropriately CorpSec’d."

A company-wide email from Gadde on January 6 announced that three tweets from Trump had been bounced, and signaled that there was a determination to find legitimate reasons to permanently suspend his account.

"Hi team, An update on actions we’re taking as a result of what’s happening in DC today. We have removed (bounced) 3 tweets from @realDonaldTrump on 6 January 2020 [sic] for making unfounded claims of voter fraud and election theft, which must be viewed in the context of the violence in Washington DC today and reasonably can be expected to continue to incite and inspire violent acts around the country."

The email continued: "We are requiring a 12-hour timeout following the removal of these tweets. We also have tweeted so there’s transparency in our actions and making it clear that future violations of the Twitter Rules will result in permanent suspension."

At 3:01 pm EST, Trump tweeted, "These are the things and events that happen when a sacred landslide election victory is so unceremoniously & viciously stripped away from great patriots who have been badly & unfairly treated for so long. Go home with love & in peace. Remember this day forever!"

In response to Trump’s tweet, Patrick Condon wrote, "what the actual f*ck," later adding, "sorry, I actually got emotionally angry seeing that. Turns out I’m not a full robot. Who know?"

Another employee responded, "it’s gut-wrenching; he’s a horrible human being."

Taibbi also revealed that Roth found Trump had "a TON of duplicate" bot applications on his account.

"By the end of the first day, the top execs are still trying to apply rules. By the next day, they will contemplate a major change in approach," wrote Taibbi.

By January 8, Taibbi wrote that "Twitter will be receiving plaudits from 'our partners' in Washington, and the sitting U.S. president will no longer be heard on the platform."

This is a breaking story and will be updated.