Everyone thinks they get a bad deal on Twitter. Both the ‘Left’ and the ‘Right’ – or more correctly, both dissenters and defenders of powerful ruling parties – feel like they are subverted by unfair suspensions, “shadow-bans” and Silicon Valley’s bias.
Of course, it’s most true right now about Donald Trump. This past week, his accounts were suspended on Facebook, Instagram and Snapchat. Parler, the MAGA-friendly app his followers are hoping to switch to, was dropped from the Google Play Store, the Apple App Store, and finally from its Amazon web hosting service.
The reason the outgoing US president has been locked out of his echo-chambers is, as Twitter said, “the risk of further incitement of violence”: note “further”, as if coaxing a mob to storm the US Capital wasn’t enough incitement.
A better explanation, as many (on Twitter) have said, is that tech giants only acted once Trump was decisively, comprehensively a lame duck. It was a hard pivot to appease Democrats, who will soon control the presidency and both houses of Congress – and therefore chair the congressional committees that are whetting their knives to cut Big Tech down to size.
So it wasn’t done on principle. If it had been, it would have happened earlier. Or else, it would have happened in India, where leaders and opinion-makers are constantly priming their followers for violence – or even for a society where violence is the rule against certain kinds of people. They only have to use a bit of code to name those people; nowadays they call them “peacefuls”. A smirking euphemism is all it takes to sail through Twitter’s filter, and keep the hate-speech flowing.
Still, everyone thinks they get a bad deal on Twitter and Facebook – which is exactly how the companies want it.
Social media bias
Today, outraged Trumpists in the US are attempting to protest-march in front of Twitter’s gates over to their own app Parler. A year ago, dissent-minded Indians were protest marching to Mastodon: they were outraged by Twitter’s repeat suspension of the advocate Sanjay Hegde, first for posting a revered 1936 photo of a lone dissenter in a German crowd refusing to give the Hitler salute; then for sharing a revolutionary Hindi poem, ‘Usko Phaansi Do’, adding only the translation of the title: ‘Hang Him‘.
The second time, Twitter offered to restore Hegde’s account if he deleted the tweet, but he refused, standing his ground on free speech. For his friends, followers and allies on Twitter (which include me) this was intolerable censorship. “Appears that @Twitter has finally joined the BJP IT cell,” we raged. In fact, it was just an especially sticky error in the moderation system. Hegde’s suspension was the result of very appropriate rules (no Nazi propaganda; no incitement to violence) inappropriately applied. We should have recognised that.
With every political camp out crying “don’t censor us”, Jack and Zuck are only happy to oblige. It means fewer costs and fewer of the subtle, thankless responsibilities of moderating content. It also means more outrage, more virality, more traffic, more data, alongside more (coded) calls for terrorism, insurrection, or mass violence – but oh well. We don’t want to risk taking down the wrong post.
It’s pretty clear that social media firms aren’t censoring for ideology. If they have to favour governments, they do so by failing to apply community standards – not by exercising them. Twitter and Facebook aren’t overactive against anyone; rather they’ve long failed to act against the violent discourse of the far right, because it’s the discourse favoured by people in power. In the US, that just changed; in India, it hasn’t.
This crude negotiation with ruling party ideology takes place in bright daylight. In February 2019, before the Lok Sabha election, the chair of India’s parliamentary committee on IT tried to drag Twitter’s CEO, Jack Dorsey, to New Delhi to be grilled on why far-right Indian accounts were being restricted. (One year later, at a political rally, this MP was calling out, “Desh ke gaddaron ko –” so the crowd could scream back, “Goli maaro saalon ko”.) Jack never came.
By wealth, distance and jurisdiction, Bay Area tech firms are insulated from the control of national governments. They are better positioned to referee free speech than, in today’s India, any agency or pillar of government. It’s a kind of compliment to Twitter that India’s far-right is so aggrieved with it; at least as aggrieved as the “Left”. You couldn’t say as much about India’s courts.
The signal, long-view danger that the social media giants face is of being blocked themselves, as many are by the People’s Republic of China. We are hearing more, and more persuasive, arguments for home-grown interests to displace Silicon Valley – “nationalist” platforms that would reverse ‘digital colonialism’ and stop the drain of data-dividends out of India.
Need for content moderation reform
Unlike Twitter and Facebook, these platforms would probably offer full compliance to the government – ending our only effective recourse against political manipulation and extremism. The day may come when we look back fondly on the role of Jack and Zuckerberg in mediating our political speech.
Until then, the best case isn’t one for fewer restrictions by social media firms, but more.
It’s a bargain fraught with risk, of course, to summon these mercenary forces and ask them to serve democracy – the genie tends to twist the wish. Yet in nearly every other medium, from conversing in person to writing in the newspaper, we have evolved civil bounds on how we speak; mostly on the manner of our speech and not the content.
We need these bounds on our social media as well. (There should be a TV show where prominent Indians are challenged to say aloud the things they say on Twitter.) Ultimately, anything that needs to be said can be said in a way that doesn’t flirt with hate speech.
We have to offer our collective permission to social media giants to moderate speech at any end of the spectrum – and to err on the side of acting. Until we do so, we allow them to prevaricate, pretending to weigh lofty principles, while their websites breed murderous discourse and genocidal attitudes toward vulnerable people (whose own freedoms, in flesh-and-blood and not just on Twitter, are in jeopardy because of this discourse).
Of course, their systems need changes, as well as much more use. Twitter users deserve a channel of recourse, such as human ombudsmen, for when errors occur (especially for regular users; public figures often have the clout to reverse a bad call, as when Salil Tripathi, a writer and free-speech advocate, was mistakenly suspended last month.)
It’s also a losing battle to moderate the fringe while you’re helping create it: If your integral design and targeting algorithms are pushing users’ opinions to new extremes.
Finally, if we’re still shying away from more moderation, it’s also because we’re judging current systems by the current state of online speech. We’ve ignored the huge, constitutive role that Twitter and Facebook now play in shaping the overall qualities of speech in societies – and enabling its worst developments. If people find “community standards” an outrageous idea, in 2021, it’s largely because social media has conditioned them away from civil speech for over a decade.
The greatest danger isn’t that we over-moderate how we speak to each other today – it’s that we fail to moderate how we will speak to each other tomorrow. Lax hate speech enforcement only helps radicalise societies into places where more and more speech is hate speech, and the cost of moderating it is higher and higher. Failing to enforce standards today, Facebook only makes its own job harder tomorrow. “The risk of further incitement” doesn’t end with Donald Trump.