The Contradictions of Twitter’s ‘We Care’ Campaign

With bots, fake news, and an angry right-wing flank, Twitter is crashing against the limits of ideological neutrality.

Jacob Silverman | Longreads | July 2018 | 8 minutes (1,917 words)

Last week, James Woods, an actor who supports Trump and delights in right-wing memes, tweeted a photo of a woman at a rally. The woman wore a yellow placard that read “My Legs Are Open For Refugees.” Woods, amused and apparently unimpressed by the woman’s appearance, wrote, “Finally, a real solution to stop illegal immigration.” The photo was fake. The words “Legs Are” were, clearly—and quite poorly—edited into the image. That didn’t stop Woods from sharing the photo, which was retweeted more than 7,400 times and picked up by Nigel Farage, a prominent member of the UK Independence Party, and popular pro-Trump accounts. For a certain faction of right-wing Twitter users, the image was another piece of evidence validating a noxious truth.

Eventually, a photographer named Lasia Kretzel saw the picture. Kretzel was disturbed; she had taken the original photograph, in which the sign read, “My Door Is Open for Refugees.” The sign also had a stamp from Amnesty International that had been removed in some edits. Kretzel had captured the image more than two years ago, at a demonstration supporting Syrian refugees in Saskatoon, Canada. In a series of tweets, she laid out how it had been altered as a piece of propaganda designed to whip up anti-refugee sentiment. Her thread was retweeted more than 6,400 times. Soon, Farage deleted his wrongful tweet. But the Woods post remained up, zooming past Kretzel’s in popularity. It’s likely that few, if any, of his followers ever noticed the correction. Besides, even if the tweet was untrue, it was still a good troll, and it showed his fans what they wanted to see. 

Twitter has no answer to these kinds of outrages. Kretzel’s methodical takedown of a blatant falsehood was left to compete, unevenly, with the splenetic tweets of a racist character actor. Twitter—“the free speech wing of the free speech party,” according to an early unofficial mantra—has given people wide latitude to post almost whatever they want. Users are made to distinguish right from wrong. The result now is that Twitter has become both a feverish global newswire and a readymade source of abusive and white supremacist content.

Instead of taking a hard look at its most volatile users, Twitter has reached out to them, holding a series of meetings with prominent conservatives (Grover Norquist, Greta van Susteren, Sean Hannity, Ted Cruz). At these gatherings, according to The Washington Post, the Republicans have groused about Silicon Valley left-wing biases and pleaded for more ideological diversity while Jack Dorsey, Twitter’s CEO, has promised to take their concerns under advisement. He has not addressed the conspiracies and falsehoods spread by people like Sean Hannity—a purveyor of the lie that Seth Rich, a staffer at the Democratic National Committee, was murdered for leaking material to Wikileaks. Such appeasements have earned Twitter a good deal of bad press, user discontent, and a reputation for being far too tolerant of Nazis. (Twitter executives have said that the verification system is broken, granting endorsement in place of an identity check, but they have not presented a solution.) Early this month, Seth Rogen, the comedian, revealed that he had been exchanging private messages with Dorsey about the practice of granting verified status to white supremacists. Rogen’s conclusion? “The dude simply does not seem to give a fuck.”

***

Lately, however, Twitter seems desperate to show that it does care. In March, Dorsey posted a thread “committing Twitter to help increase the collective health, openness, and civility of public conversation, and to hold ourselves publicly accountable towards progress.” His message included a number of mea culpas. “We have witnessed abuse, harassment, troll armies, manipulation through bots and human-coordination, misinformation campaigns, and increasingly divisive echo chambers,” he admitted. “We aren’t proud of how people have taken advantage of our service, or our inability to address it fast enough.”

In his pronouncement, Dorsey used the word “health” seven times, saying that his company wanted to find a quantitative way to measure users’ wellbeing. To that end, Twitter is partnering with an organization called Cortico, which has identified four “indicators” of social media health: shared attention, shared reality, variety of opinion, and receptivity. A Cortico blog post offered one-line definitions for each of these indicators—“shared reality,” for example, refers to whether people are using “the same facts”—but called them prototypes, open to change, and it’s unclear how these measures will be put toward strengthening the Twitter public sphere. (Deb Roy, Cortico’s co-founder and chairman, didn’t respond to requests for comment.)

Twitter has become both a feverish global newswire and a readymade source of abusive and white supremacist content.

So far, Twitter’s main contribution to improving the health of its platform has been a program to cull fake and automated accounts that have been used to spread misinformation, juice follower counts, or otherwise cause trouble. A study last year by researchers at the University of Indiana and the University of Southern California estimated that between 9% and 15% of active Twitter accounts are bots; Twitter’s own pre-cull estimation was that the cleanup campaign would slash up to 6 percent of all accounts. Recent efforts in the bot-expulsion project have identified about 9.9 million “potentially spammy or automated accounts per week,” and have removed 214 percent more accounts than Twitter did the previous year. (To date, most of the big losses were seen by celebrities—Justin Bieber lost about 3 million followers, President Trump saw a decline of about 200,000—and the average Twitter user won’t notice a thing.) The company is also hiding tweets from users deemed to be disruptive. Twitter-connected apps have received scrutiny, too, with the company suspending 142,000 of them. Spam reports are down, Twitter says, meaning that people are seeing less junk in their feeds. According to the company’s blog: “We’re preventing or catching more of this activity ourselves before you ever see it on Twitter.”

***

But how can the “health” of Twitter be adequately measured? However pious in its intentions, Twitter risks applying a technocratic sensibility to deeply-rooted social and cultural concerns that resist easy categorization. Trying to measure something nebulous, like conversational health, neglects more obvious problems, such as personal abuse and white supremacist material. Purging fake accounts may do away with some misinformation and spam, yet there is another, less-acknowledged challenge to curing what ails Twitter: a rotten culture.

Underlying Twitter’s apparent reluctance to perform intense moderation—from removing popular false rumors to banning Nazis—is the market imperative of constant growth. When The Washington Post reported on the purge of fake accounts, Twitter’s stock declined as much as 8 percent. Ned Segal, the company’s CFO, later tweeted that most of the lost accounts were less than 30 days old and were therefore not included in the company’s measure of monthly active users; the stock price recovered. But the episode revealed a fundamental tension: Any apparent decline in the number of users imperils the company’s valuation, even if the deletion of accounts is done to improve user experience. The interests of Twitter’s two main constituents, users and investors, seem hopelessly divergent.

Writing for Slate, Will Oremus wondered recently why Twitter’s stepped-up efforts to police its platform have yielded such mixed results, especially in the theater of reputation. Given the new array of enforcement programs, and Dorsey’s efforts at transparency, Oremus argued, “the company’s critics may be forced to acknowledge that the main problem is no longer a lack of effort or ingenuity on Twitter’s part. The main problem is Twitter’s fundamental structure”—a public platform that facilitates anonymity. The price of a global public square, he wrote, may be “a certain level of sketchiness.”

“Sketchiness,” like “health,” is subjective. But this square has no barricades, and no police to shoo you along if, like Alex Jones of InfoWars, you start shouting that a mass shooting was faked. In a recent statement to the House Judiciary Committee, Nick Pickles, Twitter’s senior strategist for public policy, pledged that his platform’s rules were based on standards of behavior, not ideology. “We have deliberately taken this approach as a robust defense against bias, as it requires us to define and act upon bad conduct, not a specific type of speech,” he said. “Our purpose is to serve the conversation, not to make value judgments on personal beliefs.” But this deliberate disinterest in the content of speech can be seen as a form of malign neglect. In order to provide a healthy atmosphere for its users, Twitter needs to make fine-grained, essential distinctions about when inflammatory speech turns to hate speech; it must parse the sometimes murky difference between truth and lies (at scale, no less). The onslaught of misinformation surrounding the 2016 election showed how unready social media companies were, and are, to meet this challenge. Striving to seem impartial, they have failed to reckon with the reality that some of their most prominent users—including the president—deal in bad faith; they lie at will, while accusing their opponents of purveying fake news. Such users resist automated moderation or simple fact-checking, and they will forever be the martyrs of their own narratives, their free speech threatened by what they see as the thin-skinned lefties of Silicon Valley. But they must be dealt with all the same.

***

Dorsey’s recent outreach to conservative mandarins inspires little confidence. And as he made overtures to the right-wing, a similar effort was made by Facebook, which will, in a project led by Jon Kyl, a Republican former senator, review the news feed for ideological bias. (Kyl, of course, has his own loyalties: among other ventures, he’s expected to guide Brett Kavanaugh through the Supreme Court nomination process.) What’s troubling, in both circumstances, is that no matter the outcome, conservatives win. Either they get to claim victimhood—according to which the media, Hollywood, and Silicon Valley are prejudiced against them—or they gain concessions ensuring that their points of view (and their products) are well represented. Preaching the bias of social media platforms has generated publicity, congressional subcommittee hearings, and an official Facebook apology to Lynnette Hardaway and Rochelle Richardson, a duo of pro-Trump internet commentators known as Diamond & Silk. In a bizarre twist, some Republicans have begun speaking favorably of antitrust action against big tech companies; during an April interview with Breitbart, Ted Cruz suggested using antitrust law to rein in Facebook, which he said has been “suppressing the views of conservatives.”

Purging fake accounts may do away with some misinformation and spam, yet there is another, less-acknowledged challenge to curing what ails Twitter: a rotten culture.

At the same time, embracing conspiracy-mongering has left Republicans only a few degrees removed from the likes of InfoWars—Donald Trump regularly tweets disproven theories about everything from the FBI’s handling of the DNC’s servers to the threat posed by MS-13; at a recent House Judiciary Committee hearing, Congressman Steve King, who has retweeted neo-Nazis, asked a Facebook representative why The Gateway Pundit, a site that spreads racist lies in the aftermath of terrorist attacks, doesn’t get more traffic. Despite all this, Republicans are still allowed to present themselves as the ideological mirror of Democrats—and Twitter and Facebook haven’t been willing to alienate those audiences.

With bots, fake news, and the delicate act of policing its angry right-wing flank, Twitter, along with the rest of Silicon Valley, is crashing against the limits of ideological neutrality. Whether the company likes it or not, social media platforms are publishers with beliefs and biases, some of which are political. The sooner Twitter acknowledges that, in its foundering attempts toward equivocation, it has enabled some of the vilest fringe personalities to colonize its service, the better. The obvious solution is to declare, in terms more specific than promoting “healthy conversation,” just what the company believes—and take responsible for the messes it has made.

***

Jacob Silverman is the author of Terms of Service: Social Media and the Price of Constant Connection.

***

Editor: Betsy Morais
Fact-checker: Ethan Chiel