It’s January 4th, 2023, which means Twitter Files stories have been coming out for over a month. Because these are weedsy tales, and may be hard to follow if you haven’t from the beginning, I’ve written up capsule summaries of each of the threads by all of the Twitter Files reporters, and added links to the threads and accounts of each. At the end, in response to some readers (especially foreign ones) who’ve found some of the alphabet-soup government agency names confusing, I’ve included a brief glossary of terms to help as well.
In order, the Twitter Files threads:
TWITTER AND THE HUNTER BIDEN LAPTOP STORY
Recounting the internal drama at Twitter surrounding the decision to block access to a New York Post exposé on Hunter Biden in October, 2020.
Key revelations: Twitter blocked the story on the basis of its “hacked materials” policy, but executives internally knew the decision was problematic. “Can we truthfully claim that this is part of the policy?” is how comms official Brandon Borrman put it. Also: when a Twitter contractor polls members of Congress about the decision, they hear Democratic members want more moderation, not less, and “the First Amendment isn’t absolute.”
THE “EXITING” OF TWITTER DEPUTY GENERAL COUNSEL JIM BAKER
A second round of Twitter Files releases was delayed, as new addition Bari Weiss discovers former FBI General Counsel and Twitter Deputy General Counsel Jim Baker was reviewing the first batches of Twitter Files documents, whose delivery to reporters had slowed.
TWITTER’S SECRET BLACKLISTS
Bari Weiss gives a long-awaited answer to the question, “Was Twitter shadow-banning people?” It did, only the company calls it “visibility filtering.” Twitter also had a separate, higher council called SIP-PES that decided cases for high-visibility, controversial accounts.
Key revelations: Twitter had a huge toolbox for controlling the visibility of any user, including a “Search Blacklist” (for Dan Bongino), a “Trends Blacklist” for Stanford’s Dr. Jay Bhattacharya, and a “Do Not Amplify” setting for conservative activist Charlie Kirk. Weiss quotes a Twitter employee: “Think about visibility filtering as being a way for us to suppress what people see to different levels. It’s a very powerful tool.” With help from @abigailshrier, @shellenbergermd, @nelliebowles, and @isaacgrafstein.
THE REMOVAL OF DONALD TRUMP, October 2020 - January 6th, 2021
First in a three-part series looking at how Twitter came to the decision to suspend Donald Trump. The idea behind the series is to show how all of Twitter’s “visibility filtering” tools were on display and deployed after January 6th, 2021. Key Revelations: Trust and Safety chief Yoel Roth not only met regularly with the FBI and the Department of Homeland Security, but with the Office of the Director of National Intelligence (ODNI). Also, Twitter was aggressively applying “visibility filtering” tools to Trump well before the election.
THE REMOVAL OF DONALD TRUMP, January 7th, 2021
This thread by Michael Shellenberger looks at the key day after the J6 riots and before Trump would ultimately be banned from Twitter on January 8th, showing how Twitter internally reconfigured its rules to make a Trump ban fit their policies.
Key revelations: at least one Twitter employee worried about a “slippery slope” in which “an online platform CEO with a global presence… can gatekeep speech for the entire world,” only to be shot down. Also, chief censor Roth argues for a ban on congressman Matt Gaetz even though it “doesn’t quite fit anywhere (duh),” and Twitter changed its “public interest policy” to clear a path for Trump’s removal.
THE REMOVAL OF DONALD TRUMP, January 8th, 2021
As angry as many inside Twitter were with Donald Trump after the January 6th Capitol riots, staffers struggled to suspend his account, saying things like, “I think we’d have a hard time saying this is incitement.” As documented by Weiss, they found a way to pull the trigger anyway.
Key revelations: there were dissenters in the company (“Maybe because I am from China,” said one employee, “I deeply understand how censorship can destroy the public conversation”), but are overruled by senior executives like Vijaya Gadde and Roth, who noted many on Twitter’s staff were citing the “Banality of Evil,” and comparing those who favored sticking to a strict legalistic interpretation of Twitter’s rules — i.e. keep Trump, who had “no violation” — to “Nazis following orders.”
TWITTER, THE FBI SUBSIDIARY
Twitter’s contact with the FBI was “constant and pervasive,” as FBI personnel, mainly in the San Francisco field office, regularly sent lists of “reports” to Twitter, often about Americans with low follower counts making joke tweets. Tweeters on both the left and the right were affected.
Key revelations: A senior Twitter executive reports, “FBI was adamant no impediments to sharing” classified information exist. Twitter also agreed to “bounce” content on the recommendations of a wide array of governmental and quasi-governmental actors, from the FBI to the Homeland Security agency CISA to Stanford’s Election Integrity Project to state governments. The company one day received so many moderation requests from the FBI, an executive congratulated staffers at the end for completing the “monumental undertaking.”
THE FBI AND HUNTER BIDEN’S LAPTOP
The Twitter Files story increases its focus on the company’s relationship to federal law enforcement and intelligence, and shows intense communication between the FBI and Twitter just before the release of the Post’s Hunter Biden story.
Key Revelations: San Francisco agent Elvis Chan “sends 10 documents to Twitter’s then-Head of Site Integrity, Yoel Roth, through Teleporter, a one-way communications channel from the FBI to Twitter,” the evening before the release of the Post story. Also, Baker in an email explains Twitter was compensated for “processing requests” by the FBI, saying “I am happy to report we have collected $3,415,323 since October 2019!”
HOW TWITTER QUIETLY AIDED THE PENTAGON’S COVERT ONLINE PSYOP CAMPAIGN
Lee Fang takes a fascinating detour, looking at how Twitter for years approved and supported Pentagon-backed covert operations. Noting the company explicitly testified to Congress that it didn’t allow such behavior, the platform nonetheless was a clear partner in state-backed programs involving fake accounts.
Key revelations: after the U.S. Central Command (CENTCOM) sent over a list of 52 Arab-language accounts “we use to amplify certain messages,” Twitter agreed to “whitelist” them. Ultimately the program would be outed in the Washington Post in 2022 — two years after Twitter and other platforms stopped assisting — but contrary to what came out in those reports, Twitter knew about and/or assisted in these programs for at least three years, from 2017-2020.
Lee wrote a companion piece for the Intercept here:
TWITTER AND “OTHER GOVERNMENT AGENCIES”
The Christmas Eve thread (I should have waited a few days to publish!) further details how the channels of communication between the federal government and Twitter operated, and reveals that Twitter directly or indirectly received lists of flagged content from “Other Government Agencies,” i.e. the CIA.
Key revelations: CIA officials attended at least one conference with Twitter in the summer of 2020, and companies like Twitter and Facebook received “OGA briefings,” at their regular “industry” meetings held in conjunction with the FBI and the Department of Homeland Security. The FBI and the “Foreign Influence Task Force” met regularly “not just with Twitter, but with Yahoo!, Twitch, Cloudfare, LinkedIn, even Wikimedia.”
HOW TWITTER RIGGED THE COVID DEBATE
David Zweig drills down into how Twitter throttled down information about COVID that was true but perhaps inconvenient for public officials, “discrediting doctors and other experts who disagreed.”
Key Revelations: Zweig found memos from Twitter personnel who’d liaised with Biden administration officials who were “very angry” that Twitter had not deplatformed more accounts. White House officials for instance wanted attention on reporter Alex Berenson. Zweig also found “countless” instances of Twitter banning or labeling “misleading” accounts that were true or merely controversial. A Rhode Island physician named Andrew Bostom, for instance, was suspended for, among other things, referring to the results of a peer-reviewed study on mRNA vaccines.
HOW TWITTER LET THE INTELLIGENCE COMMUNITY IN
TWITTER AND THE FBI “BELLY BUTTON”
These two threads focus respectively on the second half of 2017, and a period stretching roughly from summer of 2020 through the present. The first describes how Twitter fell under pressure from Congress and the media to produce “material” showing a conspiracy of Russian accounts on their platform, and the second shows how Twitter tried to resist fulfilling moderation requests for the State Department, but ultimately agreed to let State and other agencies send requests through the FBI, which agent Chan calls “the belly button of the USG.” Revelations: at the close of 2017, Twitter makes a key internal decision. Outwardly, the company would claim independence and promise that content would only be removed at “our sole discretion.” The internal guidance says, in writing, that Twitter will remove accounts “identified by the U.S. intelligence community” as “identified by the U.S.. intelligence community as a state-sponsored entity conducting cyber-operations.”
The second thread shows how Twitter took in requests from everyone — Treasury, HHS, NSA, FBI, DHS, etc. — and also received personal requests from politicians like Democratic congressman Adam Schiff, who asked to have journalist Paul Sperry suspended.
GLOSSARY OF “TWITTER FILES” TERMS
Government Agencies and NGOs
CISA: The Cybersecurity and Infrastructure Security Agency, an agency within the Department of Homeland Security (DHS)
CENTCOM: Central Command of the Armed Forces
ODNI: Office of the Director of National Intelligence
FITF: Foreign Influence Task Force, a cyber-regulatory agency comprised of members of the FBI, DHS, and ODNI
“OGA”: Other Government Agency, colloquially — CIA
GEC: Global Engagement Center, an analytical division of the U.S. State Department
USIC: United States intelligence community
HSIN: Homeland Security Information Network, a portal through which states and other official bodies can send “flagged” accounts
EIP: Election Integrity Project, a cyber-laboratory based at Stanford University that sends many reports to Twitter
DFR: Digital Forensic Research lab, an outlet that performs a similar function to the EIP, only is funded by the Atlantic Council
IRA: Internet Research Agency, the infamous Russian “troll farm” headed by “Putin’s chef,” Yevgheny Prigozhin
Twitter or Industry-specific terms
PII: Can have two meanings. “Personally identifiable information” is self-explanatory, while a “Public Interest Interstitial” is a warning placed over a tweet, so that it cannot be seen. Twitter personnel even use “interstitial” as a verb, as in, “Can we interstitial that?”
JIRA: Twitter’s internal ticketing system, through which complaints rise and are decided
PV2: The system used at Twitter to view the profile of any user, to check easily if it has flags like “Trends Blacklist”
SIP-PES Site Integrity Policy — Policy Escalation Support. SIP-PES is like Twitter’s version of a moderation Supreme Court, dealing with the most high-profile, controversial rulings
SI: Site integrity. Key term that you’ll see repeately in Twitter email traffic, especially with “escalations,” i.e. tweets or content that have been reported for moderation review
CHA: Coordinated Harmful Activity
SRT: Strategic Response Team
GET: Global Escalation Team
VF: Visibility Filtering
GUANO: Tool in Twitter’s internal system that keeps a chronological record of all actions taken on an account
VIT: Very Important Tweeter. Really.
GoV: Glorificaiton of Violence
BOT: In the moderation content, an individualized heuristic attached to an account that moderates certain behavior automatically
BME: Bulk Media Exploitation
EP Abuse: Episodic abuse
PCF: Parity, commentary and fan accounts. “PCF” sometimes appears as a reason an account has escaped an automated moderation process, under a limited exception
FLC: Forced Login Challenge. Also called a “phone challenge,” it’s a way Twitter attempts to verify if an account is real or automated. “Phone challenges” are seen repeatedly in discussions about verification of suspected “Russia-linked” accounts
IO: Information Operations, as in The GEC’s mandate for offensive IO to promote American interests.
This page will be kept open and updated as needed. If you have questions about terms, please send them to email@example.com