Half the room is jumping up and down, screaming “Gotcha!”. The other half shrugs its shoulders, muttering “So what’s new?”. Welcome to the war over the so-called Twitter Files.
Over the past month, Twitter’s new owner, Elon Musk, has made available to a handpicked group of journalists internal documents and conversations that took place before his takeover. They are mainly discussions about who and what should be moderated or banned, ranging from the Hunter Biden laptop story to the question of whether to remove Donald Trump from the platform. The journalists have made public selected slices of the data through a drip feed of Twitter threads.
For some, the Twitter Files provide evidence of collusion between tech companies, liberal politicians and the “deep state” to silence conservatives. For others, they constitute little more than a publicity exercise that told us nothing that we did not know. As someone who has long made a case for the importance of free speech, I think we should take the Twitter Files seriously, but also look at the discussion with a sceptical eye, given that much of it is framed by the culture wars.
Consider the controversy over “shadow banning”. It’s a phrase much bandied about in debates about social media, but its meaning is contested. For the curators of the Twitter Files, “shadow banning” means using algorithms to “deamplify” tweets – that is, stop them reaching a wider audience.
For the old pre-Musk Twitter, however, it meant something different – “deliberately making someone’s content undiscoverable to everyone except the person who posted it”, as a 2018 blog post put it. The conflicting definitions have allowed the critics to accuse former Twitter executives of lying, although they seem as interested in putting the old Twitter regime in the dock as in uncovering the truth.
The journalists curating the Twitter Files have presented deamplification as a secret process. Certainly, the internal mechanisms of moderation have been veiled in secrecy. As a practice, however, Twitter has long been open about deamplification – or “visibility filtering” as it also calls it – noting even in its terms of service that “we may… limit distribution or visibility of any Content”.
It’s a practice championed not just by old Twitter but new Twitter, too. Shortly after buying the platform, Musk tweeted that his policy would be one of “freedom of speech, but not freedom of reach”, adding that “negative/hate tweets will be max deboosted” – in other words, deamplified.
But is it a good policy? Who decides who or what should be “deamplified”? By what criteria? And how different is this from straightforward censorship? Unfortunately, the polarisation of the debate seems to have overridden a nuanced debate around such questions.
The Twitter Files also reveal some of the processes by which users are banned. Twitter insists that it does not suspend accounts for “political reasons”. That’s hard to square with the evidence.
Take the controversy over Trump. Twitter executives held anguished discussions about how to deal with the president, recognising the dangers of banning a democratically elected leader. Nevertheless, as one executive put it, “the narrative that trump [sic] and his friends have pursued over the course of this election and frankly last 4+ years must be taken into account”. In other words, Trump’s politics mattered. After the 6 January riot at the Capitol, the pressure that Twitter faced led to the second argument taking precedence over the first.
All this has strengthened the view on the right of tech companies as fomenting a liberal conspiracy against conservatives. Within Twitter, Musk tweeted, “rules were enforced against the right, but not against the left”.
Yet a 2021 study suggested the opposite. Looking at millions of tweets in seven countries – Britain, America, Canada, France, Germany, Spain and Japan – researchers found that, with the exception of Germany, Twitter algorithms amplified rightwing politicians more than the left. They also found that in America, conservative news sources were boosted more than liberal ones.
These results cut against the grain of conventional wisdom. It may be that there’s a sharp divide between algorithmic decisions and those made by human moderators. It may also be that the high-profile censoring of rightwing voices is not representative of the mass of decisions. In the last six months of 2021 alone, Twitter censored an extraordinary 4m tweets – a figure that itself should give us pause. The political bias in those 4m decisions is unknown. There is little discussion, for instance, about the suppression of Palestinian voices, a practice that has continued under Musk.
At the very least, we need more transparency about moderation. Yet, when Alex Stamos, of Stanford University’s Internet Observatory, asked Musk to allow not just handpicked journalists but academic researchers to inspect the data, Musk dismissed him on the bizarre grounds that “you operate a propaganda platform”. Musk seems more interested in being seen to “own the libs” than in opening up Twitter’s inner workings.
The most worrying issue the Twitter Files have exposed is the level of contact between the social media company and state security organisations. The FBI regularly holds meetings with Twitter executives, pressuring them to take action against “misinformation”, even when this amounted to little more than a satirical tweet, and demanding the personal data of users. Twitter, to its credit, often pushed back. Nevertheless, the Twitter Files do show an unhealthy relationship between social media and state security.
Equally unhealthy is the response of many liberals who have become sanguine about the work of the security apparatus. There has been a remarkable partisan shift in American attitudes towards the FBI, with a huge swing in Democratic support for the organisation. Many now view the FBI as an essential weapon against populism. Many seem to have forgotten the sordid history of the FBI in undermining radical movements from unions to civil rights organisations. The insouciance of liberals and many on the left to such state interference in public life is disquieting.
Twitter, we are constantly told, is not real life. That’s true. But, like all social media, it plays an inordinately large role in real life, a private company that has become an intimate part of the global public square. We need to keep that public square as open as possible. That is why the revelations of the Twitter Files matter. And that is why we need to understand their significance beyond the clamour of the culture wars.