December 27, 2022 - 10:07am

Who would you rather have banned from Twitter, alleged Covid misinformer Alex Berenson or alleged real-time doxxing account @ElonJet? That is ultimately the question raised by the latest batch of internal Twitter documents leaked by Elon Musk’s new Twitter regime. For anyone who’s followed Twitter’s misadventures over the past year, the documents are not terribly surprising in painting a company incapable of administering the public square it had stumbled into creating. The documents are important in revealing how moderation occurs on a platform such as Twitter, though little will be surprising to those who have followed the revelations about Facebook’s moderation policies in recent years.

Yet the documents only half-succeed in advancing the narrative Musk is clearly interested in pushing. If the goal is to make the old regime look bad, the documents score a slam dunk. But if the goal is to make Musk’s regime look like an improvement, they fall short and only make the case for yanking Twitter’s authority and remit from anyone.

The most irrefutably damning accusation levelled by the Free Press’s David Zweig in examining the documents is that everyday moderation decisions on Covid-related tweets were outsourced, as such decisions often are, to wholly unqualified contractors following a decision tree itself created by unqualified Twitter employees.

Whether or not such procedures stemmed from an establishment bias, and whether or not whatever bias was present was ultimately beneficial or harmful, the more profound question is: how could anyone (not just Elon Musk) do any better? Twitter was already on record as having wrongly censored or banned accounts for tweeting accurate information, while Alex Berenson himself was able to get back on the platform after filing a lawsuit, resulting in more publicity for him.

The question of what qualifies as misinformation is far more complicated than just shoring up “establishment” narratives. A June paper in the prestigious journal Nature trumpeted that 25% of infected children wind up with “long Covid”, burying the statistic that of the grab-bag of symptoms comprising “long Covid”, the most common by far were mood symptoms comprising “sad, tense, angry, anxiety, depression”, so that a sad kid contributes to the same overall statistic as one with memory loss or loss of smell. Many credentialed experts reported the 25% figure and nothing else on Twitter. If not misinformation, it was misleading and unnecessarily alarmist. At a point you start to wonder if some of those people weren’t stealthily trying to discredit actual Covid risks by intentionally leaning on the scales.

And yet none of it may have mattered all that much. Whatever Twitter did in the way of removing information, I see little convincing evidence that it significantly shifted the public debate toward the direction of the “establishment narrative”, which wholly failed to take hold in wide pockets across the country and around the world. (And judging by the number of maskless people in the New York subway even before Musk took over Twitter, that narrative was not holding up anyway.)

That leaves the upshot of the leaks as issues of principle: what should and should not be moderated out? But the issue is not what moderation regime there should be, but that Old Twitter was incapable of competently implementing any moderation regime.

The problem for Musk is that New Twitter still isn’t. Musk’s new regime eliminated the Covid decision tree, but Covid information is no longer the hot-button issue it once was. The greater moderation problem is not going away, as Musk himself indicated when banning Kanye West for tweeting a swastika merged with a Star of David, saying “he again violated our rule against incitement to violence”. I won’t miss Kanye and his Nazi apologetics, but moderating for “incitement to violence” is not so much easier than moderating for Covid misinformation. And with Musk or his lieutenants engaging in bans of parody accounts, journalists, and (briefly) promotion of accounts on competing networks, Musk’s claim to be a free speech absolutist remains as empty as ever.

So if a laissez-faire regime isn’t happening and moderation is unworkable, what can one do? People should stop defending Twitter’s Old Regime as some imaginary bastion of competent sanity, but so should others stop seeing Musk as any real palliative. The uncomfortable issue is that any company that gets into the business of shaping public discourse is going to come under government pressure and inevitably botch the mission to one extent or another. The problem with Twitter may simply be that it is Twitter.

With populations increasingly unwilling to agree on anyone to trust as an authority on a wide variety of subjects, the very idea of a consensus to enforce is breaking down. (Even Kanye has his defenders.) The best we may be able to manage is to devolve control away from any single public sphere with too much influence, because from Twitter’s example we’ve seen that no one could manage it: not Jack Dorsey, not Parag Agrawal, and certainly not Elon Musk.


David Auerbach is an American author and former Microsoft and Google software engineer.

AuerbachKeller