The Twitter files: lessons in wading through nonsense

Claims with no objective evidence that are dressed up convincingly are dangerous in a society that does not question things critically.

On 3 December 2022, half-an-hour shy of midnight (UTC), a bloke called Matt Taibbi tweeted, ‘1. Thread: THE TWITTER FILES’. It was supposed to be the grand proclamation that signalled that the curtain-up exposing the truth behind—as Twitter’s brazen new CEO, Elon Musk, put it—‘What really happened with the Hunter Biden story suppression by Twitter … at 5pm ET.’

Following a 20-minute delay in tweeting the exposé, Mr Musk dressed things up cleverly, as he so often does, saying, ‘We’re double-checking some facts, so probably start live tweeting in about 40 mins.’ A response dropped to Mr Musk’s tweet asking if he would ‘consider releasing the censoring done regarding my trial’. The tweet was sent by Kyle Rittenhouse, the kid accused of murder after he walked into a violent crowd with guns, shot someone who came at him, claimed self-defence and was acquitted.

The exposé eventually started, almost with a whimper:

The words and tone also set the stage for the type of ‘reveal’ that would follow: claims unearthed by mysterious sources mixed with wild interpretations and sparse factual backing. However, it was more than enough noise to distract from the decaying state of Twitter under Mr Musk’s (lack of) leadership.

The third ‘Twitter Files’ tweet goes live soon after: ‘3. The “Twitter Files” tell an incredible story from inside one of the world’s largest and most influential social media platforms. It is a Frankensteinian [sic] tale of a human-built mechanism grown out the control [sic] of its designer.’ More claims, more descriptions of how earth-shattering the reveal will be, lots of adjectives with barely any facts published yet.

Six tweets in, we see something starting to resemble a story crop up: ‘6. As time progressed, however, the company was slowly forced to add … barriers [against irresponsible sharing of information]. Some of the first tools for controlling speech were designed to combat the likes of spam and financial fraudsters. 7. Slowly, over time, Twitter staff and executives began to find more and more uses for these tools. Outsiders began petitioning the company to manipulate speech as well: first a little, then more often, then constantly.’

It appears Mr Taibbi’s first (and so far only) claim is that ‘outsiders’ were petitioning Twitter, potentially controlling how it restricted spread of certain pieces of information. He follows this by pointing out that when such requests came, they were handled. Indeed they were; what else would one expect a Twitter employee to do if not handle flagged tweets in accordance with their policy? Keep in mind that up till now Mr Taibbi’s only point has been that ‘outsiders’ petitioned Twitter about Tweets—which is not that different from you and I reporting tweets except it gets checked sooner—but let us consider that anyway. Subsequently, Twitter dealt with the flagged tweets as per its policy, which is not illegal or immoral.

In all fairness, Mr Taibbi points out at this point that both US political parties sent in requests to Twitter to pull down content, but he quickly doubles back to claim that the Democrats did it more than Republicans. By Occam’s razor this could simply have been the result of fewer Republicans bothering enough to write to Twitter and more Democrats doing so, but Mr Taibbi connects this to monetary contributions made to each party by Twitter employees.

We are 12 tweets into this tirade at this point and Mr Taibbi claims, ‘12. The resulting slant in content moderation decisions is visible in the documents you’re about to read. However, it’s also the assessment of multiple current and former high-level executives.’ The first sentence is fair enough: we are being prepared to observe inclinations in content moderation towards one party over another. But is this not to be expected based on how there were more requests by one party than the other? Had the republicans requested more often, and more reasonably, perhaps we would have seen Twitter censor content they requested more often as well.

I could not track down tweets 13, 14 and 15 but the words ‘Okay, there was more throat-clearing about the process, but screw it, let’s jump forward,’ tell me Mr Taibbi skipped copying those tweets over from his notepad. So we move on to tweet number 16 that finally clarifies what on earth this tweet storm is all about:

Was Elon Musk’s grand Twitter reveal about the affairs of a private citizen?

Nevertheless, we proceed on the reasoning that this private citizen was in a unique position, being the son of a former Vice-President of the US. The story in question was an NY Post exposé titled ‘Biden secret e-mails’ referring to Hunter Biden but sounding like click bait because it appeared to refer to Joe Biden, who is the first ‘Biden’ of whom the general public would have thought upon reading that headline1.

‘Twitter took extraordinary steps to suppress the story,’ writes Mr Taibbi in his eighteenth tweet, ‘removing links and posting warnings that it may be “unsafe.” They even blocked its transmission via direct message, a tool hitherto reserved for extreme cases, e.g. child pornography.’ The mention of child pornography is interesting because it highlights the extent to which Twitter seemingly went to prevent this story from being shared. Yet, as with everything, these are simply claims not backed by documents. If these documents do exist, releasing them as evidence of claims is critical; simply throwing claims around while keeping supporting evidence hidden away is simply counterproductive if the primary intention is putting out the truth.

At this point an official from the US President’s office writes to Twitter enquiring why a White House spokesperson was blocked for talking about the NY Post story. Clearly, according to Twitter’s own policy, talking about a hack is allowed. Following this, Twitter’s Public Policy head2 asked the relevant department to look into the issue saying, ‘Hey team! Are you able to take a closer look here? Thank you!’ To anyone, this would be a routine message asking a team to look into an issue, but not to Mr Taibbi who interprets it as ‘a polite WTF query.’ Anyway, the response from the team in charge was that the account was blocked on the basis of it violating the company’s ‘hacked materials’ policy.

This explanation does not make sense if the White House spokesperson did in fact only speak of the news of the hack and not refer to Hunter Biden’s private information obtained as a result of this hack. Mr Taibbi assumes silently that this is the case. However we do not know for a fact what the White House spokesperson said that called for their account being blocked. If they talked, for instance, about some damning piece of news involving Hunter Biden’s personal life that was obtained as a result of the hack, they were in fact violating Twitter’s policy. At the end of the day, nobody knows.

Mr Taibbi also conveniently left out that—just as he had mentioned how Twitter was predominantly democrat-leaning—the company digging up dirt against them, NetChoice, was even more blatantly Republican-leaning.

None of this stops Mr Taibbi from jumping the gun for the third time: ‘The decision was made at the highest levels of the company, but without the knowledge of CEO Jack Dorsey, with former head of legal, policy and trust Vijaya Gadde playing a key role.’ None of these claims are backed by any evidence. Neither are the next couple of sentences: ‘“They just freelanced it,” is how one former employee characterized the decision. “Hacking was the excuse, but within a few hours, pretty much everyone realized that wasn’t going to hold. But no one had the guts to reverse it.“’ Who is this former employee who spoke about this so secretly? Whom did they speak to and when? Do they even exist? With a claim and no evidence to back it, all of what Mr Taibbi says at this point could just be entirely made up.

The White House spokesperson’s account was not, in fact, brought down. The circulation of one specific tweet was stopped—as Mr Taibbi himself shows a few tweets later—with Twitter agreeing to lift any form of shadow ban once the tweet in question was deleted. Their reasoning—again, as Mr Taibbi himself shows us—was that the news was raw and investigation was warranted; but, since there was no way to ascertain policy violations until the news was confirmed, it would be better to err on the side of caution and restrict circulation of a tweet rather than allow rampant sharing that could result in a flood of misinformation the likes of which the internet had witnessed prior to the 2015 US elections.

Like his boss Mr Taibbi dresses this up: once he points out that ‘former Deputy General Counsel Jim Baker again seems to advise staying the non-course, because “caution is warranted”’ but also colours the facts with his opinion saying Twitter’s response ‘was essentially to err on the side of… continuing to err,’ making it look like Twitter was wrong all the while.

Then comes the big swing. ‘34. NetChoice lets Twitter know a “blood bath” awaits in upcoming Hill hearings,’ writes Mr Taibbi, ‘with members saying it’s a “tipping point,” complaining tech has “grown so big that they can’t even regulate themselves, so government may need to intervene.”’

This sounds alarming. Surely if NetChoice, a company working to ‘make the Internet safe for free enterprise and free expression,’3 is saying it, the rumour must have some credible origin. But there is none. Neither did NetChoice ever disclose the facts or produce statistics, nor did Mr Taibbi ever bother to verify any himself.

Additionally, what Mr Taibbi conveniently left out of his statements was that—just as he had mentioned how Twitter was predominantly democrat-leaning—the company digging up dirt against them, NetChoice, was even more blatantly Republican-leaning. The same source that reports Twitter’s allegiances reports NetChoice as showing 100% Republican support since 20104.

Concluding the poor exposé Mr Taibbi discusses a letter from the VP of NetChoice, Carl Szabo—who has donated exclusively to the Republican Party since 2017)—who claims Democrats had no regard for the First Amendment. Keep in mind, once again, the chain of evidence is absent here. This is Mr Taibbi’s claim based on a letter purportedly from Mr Szabo to Twitter based on supposed informal discussions with lawmakers. It is, quite simply, a chain of hearsay.

In the end, the Twitter Files were nothing. As Charle Warzel rightly described them, writing for The Atlantic, the whole charade was ‘sloppy, anecdotal, devoid of context, and, well, old news.’ It served to direct public attention away from the failures of Musk’s Twitter by calling undeserved attention to old stories of old Twitter dug up on nobody’s request and dressed up as social service through sub-standard journalism.

Content moderation on social media is a complex, nuanced topic. The freedom of speech, like all freedoms, loses meaning without boundaries. Despite being a private company, having such a public voice means Twitter has a responsibility to enforce these boundaries with the only requirement being that the boundaries are objective, not subjective. And everything Mr Taibbi himself said makes it look like Twitter was being objective. You could of course throw in words like ‘secretive’ and ‘chilling’ to make things spicy (as poor journalistic standards necessitate) but none of that changes the facts of the case. As far as the facts are concerned, the Twitter Files have done nothing to put a dent in any debate.

Who better to highlight the importance of all this nonsense than the new CEO of Twitter himself? Except that Mr Musk admitted on a Twitter Spaces session (an entire day after he and his social media platform made a big hullaballoo about the exposé) that he had himself never read the infamous Twitter Files.

  1. If I said, for example, ‘Schumacher won the F1 at Japan,’ most people who are not into Formula 1 would think of Michael Schumacher, not his son Mick Schumacher. 

  2. Twitter’s public policy head, Caroline Strom, left Twitter after Mr Musk took over

  3. Mr Taibbi wrongly describes NetChoice as a ‘research firm’. 

  4. Except in 2014 when they were about 93% Republican.]