- Technology

Getting rid of QAnon won’t be as easy as Twitter might think

Every tech platform has two policies about what they will allow: the policy that’s written, and the policy that’s enforced. Ideally there would be no gap between these, but in practice it almost can’t be helped.

To take one dumb example, the embryonic audio social network Clubhouse posted its first set of community guidelines the other week. One of these rules says “we prohibit the spread of or attempts to spread false information or news.” In theory, this means that only the truth is allowed on Clubhouse. Not a bad policy, really, but when you’re building a network that someday anyone will be able to use to essentially broadcast their phone calls, I imagine that Clubhouse will find it extremely hard to enforce. Of course people will lie on Clubhouse. They will lie all the time. The question is who will be punished for doing so.

I thought about all this on Tuesday night after reading that Twitter had banned 7,000 accounts and put restrictions on 150,000 more related to the sprawling conspiracy theory slash alternate reality game slash religion QAnon.

Twitter will stop recommending accounts and content related to QAnon, including material in email and follow recommendations, and it will take steps to limit circulation of content in features like trends and search. The action will affect about 150,000 accounts, said a spokesperson, who asked to remain unnamed because of concerns about the targeted harassment of social media employees.

The sweeping enforcement action will ban QAnon-related terms from appearing in trending topics and the platform’s search feature, ban known QAnon-related URLs and prohibit “swarming” of people who are baselessly targeted by coordinated harassment campaigns pushed by QAnon followers.

The ban on “swarming” is particularly significant. Often conspiracy theorists will falsely declare various celebrities and politicians as being at the heart of some lunatic Satan-worshiping scheme, and their followers will descend on the timelines of the accused — and their friends and family — to wreak havoc.

And yet, how will swarming be detected? Twitter updated its policy by tweeting it, which means the update itself was brief to the point of being opaque. You will be suspending if you are found to be “coordinating abuse around individual victims.” OK. But what if you’re just one of thousands of accounts in the victim’s mentions tweeting “stop eating babies”? I don’t see anything in the policy about that — perhaps because, from an enforcement, perspective, it could be difficult to determine what makes abuse “coordinated.” As Evelyn Douek notes here, Twitter itself has not yet defined what it means by coordination.

In any case, it seems that Twitter won’t be the only network to take a harder stance against QAnon. Facebook plans to escalate its own enforcement against its adherents, Kate Conger reports at the New York Times. She also offers a good concise description of what QAnon is, for those still unaware:

Facebook is preparing to take similar steps to limit the reach of QAnon content on its platform, said two Facebook employees with knowledge of the plans, who spoke on the condition of anonymity. The company has been coordinating with Twitter and other social media companies and plans to make and announcement next month, the employees said. Facebook declined to comment.

The QAnon theories stem from an anonymous person or group of people who use the name “Q” and claim to have access to government secrets that reveal a plot against President Trump and his supporters. That supposedly classified information was initially posted on message boards before spreading to mainstream internet platforms and has led to significant online harassment as well as physical violence.

All of this is welcome news, since QAnon conspiracists have increasingly been associated with real-world violence. (Pizzagate, a forerunner of QAnon, famously led to a gunman shooting up a pizza restaurant.)

At the same time, QAnon is also a political movement. By some counts, there are a dozen Republicans now running for national office who profess allegiance to QAnon. Some are expected to win seats in Congress. And so will they be affected by the ban?

“A Twitter spox tells me that ‘currently candidates and elected officials will not be automatically included in many of these actions broadly,’” CNN’s Oliver Darcy reported. That suggests that some of these QAnon adherents who win Congressional seats could continue to use Twitter to promote the idea that there is a “deep state” devoted to child murder, potentially inciting their followers to violence. It’s an admittedly difficult position for Twitter (or Facebook) to be in — the bar for de-platforming an elected official should be high. And yet it’s hard to imagine that members of the QAnon caucus won’t clear it.

It’s easy to announce some strict new policy against some violent group and get some praise. But we’ve seen enough of these adjustments recently — I’m thinking here of Facebook’s Boogaloo ban — to know that the devil is very much in the details. I believe Twitter intends to eliminate as much QAnon content from the platform as it can. But I’m less sure, looking at its flimsy tweets on the subject, how it will.


On Tuesday I wrote about the New York Times’ Kevin Roose and his tweets about the most popular links on Facebook, which have aggravated company executives by painting what they say is a misleading picture of the News Feed. I got an unusually polarized response, and wanted to share it here. The pro-Roose side, which you can get a flavor of in the responses to CNN reporter Oliver Darcy’s tweet about my story, is about what you would expect: Facebook bad, go Kevin.

Meanwhile, people in Facebook’s orbit criticized me for going too easy on Roose’s tweets. The data he’s using comes from CrowdTangle, a Facebook-owned tool for analyzing the spread of stories on social networks. But what it measures is activity on Facebook publisher pages, not links shared throughout the ecosystem overall. And the people who interact heavily with Facebook publisher pages are not typical Facebook users — they’re more partisan than average, and that distorts the picture painted in the lists, people told me. But because that picture fits reporters’ general view of Facebook, which is that it is is a powerful engine for promoting the conservative movement and its leading figures, we’ve all accepted the idea that CrowdTangle data is a good gauge of what’s popular on Facebook.

I don’t know. Open CrowdTangle. “The easiest way to keep track of what’s happening on social media,” a banner blares. Log in, and a prominent tab called “Leaderboard” links you to a list of publishing categories that you can sort by interactions or views. There is not a disclaimer when you log in that says, “just so you know, CrowdTangle data paints a misleading picture of what’s really happening on Facebook.” Instead it offers a bunch of data in attractive charts about content that very much appears to be popular. Hence the leaderboard!

Anyway, if Facebook wants to do something about the way reporters like Roose represent data on a tool that Facebook itself owns, it seems to me that they could start with the tool rather than the tweets.


Last month I wrote about how the independent developer Basecamp successfully bent Apple’s App Store policies after complaining about the company’s anticompetitive practices surrounding its own email app and certain revenue it hoped to extract from Basecamp’s email app Hey. Now an antitrust hearing has been scheduled on Monday in the House of Representatives, where Apple CEO Tim Cook is expected to testify and take questions on App Store competition.

Ahead of that, today Apple held a briefing to announce that it had commissioned a study of the commissions taken by various other e-commerce players. Economists at something called the Analysis Group were called in to look up the various cuts taken by Google, Airbnb, and other platforms, and they did, and on Wednesday they published a PDF. They were neither asked nor allowed to comment on the significance of the findings or the fairness of the App Store.

Anyway, OneZero’s Will Oremus had my thoughts exactly when he said: “I wrote in depth about the antitrust case against Apple in February. The core argument was never that the 30% rate was out of line with the industry. It’s that developers had no choice but to pay it, even while competing with Apple’s own apps that do not.”

About matthewjccom

Read All Posts By matthewjccom