Highlights from this weekend's Elon Musk "Twitter Files"

Catch up on the most recent revelations from Twitter's internal documents.

As we expected, Elon Musk continues to release Twitter’s internal documents concerning censorship and political bias. He has followed up last weekend’s report from journalist Matt Taibbi with three more releases.

I’ve summarized some of the highlights below, and provided links to the relevant Twitter threads so you can read the original reporting on your own.

Twitter Files 2: Shadowbanning

The second batch of Twitter documents was released to former New York Times editor and founder of The Free Press, Bari Weiss. This batch focuses on the practice of “shadowbanning” - when a social media company secretly blocks a user’s account, posts or comments from being visible to other users.

Weiss explains how shadowbanning worked on Twitter:

Bari Weiss @bariweiss

1. A new #TwitterFiles investigation reveals that teams of Twitter employees build blacklists, prevent disfavored tweets from trending, and actively limit the visibility of entire accounts or even trending topics—all in secret, without informing users.

Bari Weiss @bariweiss

9. “VF” refers to Twitter’s control over user visibility. It used VF to block searches of individual users; to limit the scope of a particular tweet’s discoverability; to block select users’ posts from ever appearing on the “trending” page; and from inclusion in hashtag searches.

Weiss provides four examples of accounts that were shadowbanned: Stanford medical professor Dr. Jay Bhattacharya, conservative talk radio host Dan Bongino, and conservative activists Charlie Kirk and LibsofTikTok. What these four individuals have in common is that they have significant social media followings and offer perspectives that counter liberal orthodoxy and Democratic Party positions on a wide range of issues.

The power to shadowban appears to have been used in a politically biased way to target Twitter accounts on the right side of the political spectrum.

This reporting also offers a window into the rationalizations offered by Twitter’s then-Head of Trust and Safety, Yoel Roth. Roth is quoted as saying:

Bari Weiss @bariweiss

25. Roth wrote: “The hypothesis underlying much of what we’ve implemented is that if exposure to, e.g., misinformation directly causes harm, we should use remediations that reduce exposure, and limiting the spread/virality of content is a good way to do that.”

Of course, the key question here is who decided what constitutes misinformation? At Twitter, it was a small, secretive group of influential managers and executives who were not bound by any objective policies or procedures.

Bari Weiss @bariweiss

13. But there existed a level beyond official ticketing, beyond the rank-and-file moderators following the company’s policy on paper. That is the “Site Integrity Policy, Policy Escalation Support,” known as “SIP-PES.”

Bari Weiss @bariweiss

14. This secret group included Head of Legal, Policy, and Trust (Vijaya Gadde), the Global Head of Trust & Safety (Yoel Roth), subsequent CEOs Jack Dorsey and Parag Agrawal, and others.

Bari Weiss @bariweiss

15. This is where the biggest, most politically sensitive decisions got made. “Think high follower account, controversial,” another Twitter employee told us. For these “there would be no ticket or anything.”

Twitter Files 3: Political Bias Intensifies

The third batch of Twitter documents went to the man who kicked things off last weekend, Matt Taibbi. Taibbi reports on the intensification of political bias at Twitter, which creates the context for the eventual suspension of Donald Trump in January 2021.

Taibbi explains that in the months leading up to the suspension of Trump’s account, Twitter’s censorship practices became more frequent and employees were frustrated that company policies did not enable them to do more. For example, Hollywood actor James Woods makes another appearance in these documents, expressing concern over a label that Twitter placed on a Trump tweet in October 2020. Twitter employees debate how to respond, and state clearly that they don’t have a policy justification to consider Woods in violation.

Matt Taibbi @mtaibbi

@ShellenbergerMD @bariweiss @JohnBasham @TitaniaMcGrath 37. A seemingly innocuous follow-up involved a tweet from actor @RealJamesWoods, whose ubiquitous presence in argued-over Twitter data sets is already a #TwitterFiles in-joke.

Matt Taibbi @mtaibbi

@ShellenbergerMD @bariweiss @JohnBasham @TitaniaMcGrath @RealJamesWoods 38. After Woods angrily quote-tweeted about Trump’s warning label, Twitter staff – in a preview of what ended up happening after J6 – despaired of a reason for action, but resolved to “hit him hard on future vio.”

When talking about Woods, Twitter employees state they want to “hit him hard.” But, when they speak about Democrats engaged in potential misinformation, Twitter employees sing a different tune. For example, when Hollywood comedian Elijah Daniel and Hollywood writer Rick Rosner accused the U.S. Supreme Court of trying to “steal the election” from Joe Biden, Twitter employees found no reason to be concerned.

Matt Taibbi @mtaibbi

@ShellenbergerMD @bariweiss @JohnBasham @TitaniaMcGrath @RealJamesWoods 41. Meanwhile, there are multiple instances of involving pro-Biden tweets warning Trump “may try to steal the election” that got surfaced, only to be approved by senior executives. This one, they decide, just “expresses concern that mailed ballots might not make it on time.”

Matt Taibbi @mtaibbi

@ShellenbergerMD @bariweiss @JohnBasham @TitaniaMcGrath @RealJamesWoods 42. “THAT’S UNDERSTANDABLE”: Even the hashtag #StealOurVotes – referencing a theory that a combo of Amy Coney Barrett and Trump will steal the election – is approved by Twitter brass, because it’s “understandable” and a “reference to… a US Supreme Court decision.”

Twitter Files 4: Trump’s Gone

Michael Shellenberger, author of Apocalypse Never and former independent candidate for Governor of California, reported on the fourth batch of Twitter documents. This time, we see exactly how Twitter decided to suspend Donald Trump, which was controversial across the political spectrum. World leaders like Emmanuel Macron and Angela Merkel, who rarely agreed with Trump on anything, opposed Twitter’s decision at the time.

Then-CEO of Twitter, Jack Dorsey, initially resisted the idea of suspending Trump. However, he eventually succumbed to internal pressure, leading to a five-strike system to justify the suspension of accounts.

Michael Shellenberger @ShellenbergerMD

Around 11:30 am PT, Roth DMs his colleagues with news that he is excited to share. “GUESS WHAT,” he writes. “Jack just approved repeat offender for civic integrity.” The new approach would create a system where five violations ("strikes") would result in permanent suspension.

Unfortunately for some Twitter employees, this new system didn’t give them permission to suspend Trump, which is ultimately what they wanted to do.

Michael Shellenberger @ShellenbergerMD

The colleague wants to know if the decision means Trump can finally be banned. The person asks, "does the incitement to violence aspect change that calculus?” Roth says it doesn't. "Trump continues to just have his one strike" (remaining).

Trump’s involvement in the January 6 U.S. capitol attack ultimately provided the justification for suspension based on “incitement to violence.”

Michael Shellenberger @ShellenbergerMD

Roth's colleague's query about "incitement to violence" heavily foreshadows what will happen the following day. On January 8, Twitter announces a permanent ban on Trump due to the "risk of further incitement of violence."

However, there remained no internal policy justification for suspending Trump. There was no objective standard created that could then be applied to politicians, public figures or other users. One junior employee pointed this out, but it doesn’t appear his or her concerns were responded to in the company’s Slack chat.

Michael Shellenberger @ShellenbergerMD

The *only* serious concern we found expressed within Twitter over the implications for free speech and democracy of banning Trump came from a junior person in the organization. It was tucked away in a lower-level Slack channel known as “site-integrity-auto."

Additional Twitter files will be released soon. I will do another summary later in the week if interesting insights are revealed.