Skip to main content

A shadowy influence campaign on Facebook is targeting liberal activists

A shadowy influence campaign on Facebook is targeting liberal activists


Who’s behind the latest set of attacks?

Share this story

Illustration by James Bareham / The Verge

We are 97 days away from the midterm elections, there’s an active campaign to undermine our democracy on Facebook, and no one can say for certain who’s behind it. That was the big news to emerge on Tuesday in a call that Facebook held with reporters, shortly after The New York Times broke the news that the company had detected an “ongoing political influence campaign” that led it to remove 32 pages and fake accounts from the service.

Facebook laid out its major findings in a blog post. On one hand, the number of fake accounts caught by Facebook here is relatively small. On the other, they were followed by 290,000 people. The accounts were created between March 2017 and May of this year and included such pages as “Aztlan Warriors,” “Black Elevation,” “Mindful Being,” and “Resisters.”

The pages were quite active, posting 9,500 times before they were shut down. And they ran ads: 150 of them, at a cost of $11,000. (Notably, these accounts stopped posting after Facebook implemented new disclosure requirements for advertisers.)

A post from a removed page (Facebook)
A post from a removed page (Facebook)

Most provocatively, the pages seemed to be focused in part on fomenting real-world dissent. Facebook found that the pages had created 30 events since last year, the largest of which had 1,400 people scheduled to attend. According to the Digital Forensic Research Lab, a think tank focused on preventing election interference that has a partnership with Facebook, the fake accounts exclusively targeted the American left.

Facebook said it had been compelled to disclose its findings ahead of protests connected to a “Unite the Right” rally planned for August. Some of the removed pages were planning or supporting protests for the rally, which is a sequel to an event last year that turned deadly when a man linked to neo-Nazis drove a car into a group of anti-racist protesters.

The Facebook pages were divisive — and effective, according to DFRL. “Of note, the events coordinated by — or with help from — inauthentic accounts did have a very real, organic, and engaged online community; however, the intent of the inauthentic activity appeared to be designed to catalyze the most incendiary impulses of political sentiment,” it said in a blog post.

It’s reasonable to conclude that at least some of the fake accounts were Russian in origin

Citing various linguistic quirks, the lab concludes that it is reasonable to conclude that at least some of the fake accounts were Russian in origin. It also concludes that these operations are becoming more difficult for Facebook to detect:

Their behavior differed in significant ways from the original Russian operation. Most left fewer clues to their identities behind, and appear to have taken pains not to post too much authored content. Their impact was, in general, lower, compared with the 300,000 followers amassed by Russian troll account “Black Matters.”

Information operations, like other asymmetric threats, is adaptive. These inauthentic accounts, whoever ran them, appear to have learned the lessons of 2016 and 2017, and to have taken more steps to cover their traces. This was not enough to stop Facebook finding them, but it does reveal the challenge facing open source researchers and everyday users.

As the day went on, another challenge emerged: real Americans complaining their events had been shut down unfairly. It appears that Facebook deleted events even if they had just a tangential connection to one of the authentic accounts. “The Unite the Right counter protest is not being organized by Russians,” organizer Dylan Petrohilos tweeted. “We have permits in DC, We have numerous local orgs like BLM, Resist This, and Antifascist groups working on this protest. FB deleted the event because 1 page was sketch.”

In a separate thread, an organizer named Brendan Orsinger elaborated on how the event was removed. Orsinger, who is helping to organize a Unite the Right counter-protest, had added the “Resisters” page as a host of the event to help promote it. “The Resisters page was a 20K follower social media megaphone,” he tweeted, “and it helped us reach more folk.” Unfortunately for Orsinger, the page was run by someone who was faking their identity. So when Facebook killed the Resisters page, it also killed the events that page was “hosting.” (Facebook says Resisters created the event and only then invited legitimate pages to co-host.)

You can see how tricky these issues are. You can also see how unlikely it is that we’ll untangle all of them before November 6th. Facebook deserves credit for disclosing the threats publicly, in something close to real time. Especially given that the disclosure only serves to make us more worried than we were before.

And in the meantime, the mysterious maybe-Russian agents behind the current assault have succeeded in amplifying division and doubt, online and off.


Facebook shuts off access to user data for hundreds of thousands of apps

Here’s some more Cambridge Analytica fallout:

Facebook this evening announced that it’s shutting off access to its application programming interface, the developer platform that lets app makers access user data, for hundreds of thousands of inactive apps. The company had set an August 1st deadline back in May, during its F8 developer conference, for developers and businesses to re-submit apps to an internal review, a process that involves signing new contracts around user data collection and verifying one’s authenticity.

Why We’re Sharing 3 Million Russian Troll Tweets

FiveThirtyEight is sharing 3 million tweets from 2,848 Twitter handles associated with the Internet Research Agency troll farm. They’re doing it in the hopes that lots more researchers will dig into the archives and public their findings.

Bigfoot erotica and Denver Riggleman, explained

Okay, I swear I tried very hard to ignore the fact that yesterday “Bigfoot erotica” took over Twitter and was somehow connected to a race for a House seat in Virginia. But on Tuesday, I caved and read about it, and here’s what you need to know: a man with the improbable name of Denver Riggleman is the Republican nominee in Virginia’s Fifth Congressional District. Per this Matt Yglesias piece, Riggleman is the co-author of a self-published 2006 book called Bigfoot Exterminators, Inc.: The Partially Cautionary, Mostly True Tale of Monster Hunt 2006a work about people who look for Bigfoot.

Riggleman posted some explicit photos of Bigfoot to his Instagram (???), which were subsequently shared by his Democratic opponent, drawing attention to the thriving genre of Bigfoot erotica (?!?!?!), and also the whole thing was maybe a scheme to draw attention to the fact that Riggleman has associated with white supremacists.

Anyway, sorry, that’s why Bigfoot erotica was all over Twitter.


Can Comedy Twitter survive James Gunn, Dan Harmon and a war on jokes?

Twitter has been a boon to comedians. But some jokes age much, much worse than others, and now digging up old problematic tweets has become a go-to tactic of the right wing. Julia Alexander explores that tension here. My proposed solution: allow Twitter users to set their tweets to expire. Better yet: allow us to highlight a few prized tweets that we want to keep around forever, and let us pin them to our profiles.

Facebook might start holding singing competitions among its users

It’s a proven fact that Russian actors cannot sow division in the United States while they are participating in singing competitions. To that end:

Facebook appears to be working on a talent show feature that would have users record themselves singing and then submitting their videos for critique. In the app’s code, researcher Jane Manchun Wong spotted an interface that would let users choose a popular song and then record themselves singing it.

What Happened When I Tried Talking to Twitter Abusers

Here is a person who was receiving a lot of ugly tweets and tried to engage with her attackers and found that they were not receptive to what she was saying.

4,925 Tweets: Elon Musk’s Twitter Habit, Dissected

Some fun nuggets in this graphical breakdown of Elon Musk’s tweets. He replies more than any other big tech company CEO, and he is a populist: about 41 percent of his replies are to people who have fewer than 500 followers.

Goodreads and the Crushing Weight of Literary FOMO

Angela Waterslicer has a fun piece about how she forgot to turn off Goodreads notifications for many years, and now every day she gets shamed about how few books she reads, and she talks to the Goodreads CEO about it. I also had the not-reading-books problem until last year, when I finally integrated audiobooks into my reading and managed to get up to about two books a month. It’s my favorite lifehack of the past decade.

Journalist Mocked After $22 Avocado Toast Purchase Backfires

Many people have written to The Interface asking when we are going to comment on Friend of the Newsletter Taylor Lorenz’s viral tweet in which she was disappointed by the avocado toast that she ordered through Seamless, which cost $22. Credit to Lorenz here: she managed to turn a brutal self-own into a thoughtful examination of how viral tweets bring out the very worst in Twitter, as the author is quickly subjected to a comical amount of abuse from every imaginable angle.

Say what you will about $22 toast: Taylor’s initial tweet essentially served as bait that allowed her to illuminate the bizarre viral dynamics of our current content moment in a more personal way than her stories typically allow for. Also, anyone who wants to yell about how much someone spent on toast has to turn over all of their receipts to Twitter for a complete review, I’m sorry, but that’s how it works.


Facebook is personalizing the navigation bar on its mobile app

The row of things you do not use on the bottom navigation bar inside the Facebook app will now be personalized do you, raising hopes that, someday, I may stop accidentally tapping on the Marketplace tab only to see someone in Fremont selling a refurbished fax machine for $30.

How to get the new Facebook plane reaction

Facebook accidentally launched a new reaction emoji, and that emoji is a plane.


Is Facebook evil? Everything bad about Facebook is bad for the same reason

I can’t decide if this take is nuclear or not. On one hand, it compares Facebook, unfavorably, to the Third Reich. On the other hand, its central point, that Facebook’s willful passivity in the face of most things enables its worst actors, seems hard to argue with. On a third hand, the essay is poorly constructed and argued. But we were a little low on takes over here. Anyway:

That brings us back to Facebook. It has its own grand project—to turn the human world into one big information system. This is, it goes without saying, nowhere near as terrible as the project of the thousand-year Reich. But the fundamental problem is the same: an inability to look at things from the other fellow’s point of view, a disconnect between the human reality and the grand project.

And finally...

This Instagram Account Shows How Instagram Photos Look the Same

On this day where we ponder the meaning of “coordinated inauthentic comment,” I enjoyed this look at @insta_repeat, an anonymous Instagram account that finds people taking the same picture and assembles them into a grid of shame. We think of photography as a largely original medium, and yet so often certain kinds of photos on Instagram come to feel like collectible pokémon: here’s my shot in the pit of yellow balls at the Museum of Ice Cream, here’s my arm holding up a large ice cream cone against a colorfully painted wall, etc.

Anyway, go forth and be authentic, y’all!

Talk to me

Send me: tips, comments, coordinated inauthentic content.