How to Fool Americans on Twitter

The data behind 40 of the now-suspended accounts Twitter admits were Russian trolls

One of the many ways propagandists manipulate US public discourse is by imitating Americans on Twitter. As part of their appearance in front of the House and Senate Intelligence committees earlier last week, Twitter released over 4,500 accounts it had suspended as part of its internal investigation into Russian meddling on their platform. These accounts are no longer accessible by the public, but in summer 2016 my startup, New Knowledge, had already uncovered 40 of the accounts that Twitter now acknowledges were Russian trolls, and collected 7,500 of their tweets. Our data about these accounts’ activity reveal that the trolls repeatedly created laughable caricatures of US conservatives, and lied about divisive social issues to pit Americans against each other.

Russian trolls, posing as American conservatives, tapped into the most hateful, toxic conversations in US politics, and amplified them to millions of people.

The account @USA_Gunslinger was a fake account removed by Twitter for its ties to Russian disinformation

Guns and Racism: Propaganda Goes Viral

The professional Russian trolls who ran these accounts invented fake conservative personalities that they used to build an audience, weighing in on everything from immigration to gun control to feminism. They fooled a lot of people. One high-profile account that has been widely covered in the media, @Jenn_Abrams, published tweets like this one, excusing rape while equating feminism and cancer:

By the way, inventing new kinds of rape doesn’t make any useful contribution to the world #FeminismIsCancer

“She” was so successful at fooling the public that her account had over 70,000 followers, and was cited in reporting by the USA Today, Fox, and the Washington Post.

Another fake account, @SouthLoneStar, targeted the Muslim community, while taking a few shots at Democratic politicians.

Texas Lone Star’s most popular tweet, suggesting that there was voter fraud in every county won by Hillary Clinton, was retweeted over 2,000 times.

Man PROVES software stole votes in ALL ‘Hillary won’ counties! DNC rigged elections #IndictHillary #ImWithHer -@SouthLoneStar

Other accounts avoided presidential politics, instead tweeting about race, guns, immigration, and the mainstream media.

John Davis, aka @TheFoundingSon, a fake account claiming to be a “conservative, Christian, patriot,” had nearly 42,000 followers, was routinely retweeted hundreds of times, and posted overtly racist memes — like this one drawing a comparison between Hurricane Harvey looters and undocumented immigrants.

Pamela Moore, aka @Pamela_Moore13, also claimed to be from Texas, and spent “her” time making fun of liberals and the media. For example, in this tweet she combined the logs of Antifa, the radical leftist group, and CNN.

Perhaps the strangest user in the trove of Russian-linked accounts Twitter publicly acknowledged is Jihadist Wife, aka @Jihadist2ndWife. This was a Russian troll pretending to be an American, who was in turn pretending to be an ISIS supporter — as a joke.

The account framed all as Muslims as terrorists, and then made fun of them.

The Patterns of Disinformation

When we look at the data, it’s clear that these examples fit a larger pattern of disinformation that plays on racial, religious, and cultural divisions. Most of the trolls’ tweets were pro-Trump and/or anti-Clinton, but these accounts were also consistently anti-Muslim, anti-immigrant, racist, and pro guns.

Time to Act

The ultimate goal of this disinformation campaign is destabilizing the United States. At first it they attacked our politics, and then the media. Next they’ll undermine our cultural institutions, our national security forces, and our critical infrastructure, like energy, finance, and manufacturing.

Unfortunately, when it comes to social media manipulation, tech companies have a history of missing the forrest for the trees. In 2013, ISIS exploited social media to dominate the narrative in US media. Instead of fixing the vulnerabilities on their platforms, the big Silicon Valley media companies instead decided to focus specifically on rooting out ISIS supporters. Then in 2016, Twitter and Facebook were exploited again, this time by Russia. The tech companies response is eerily similar — they’ve vowed to address foreign interference in US elections, again narrowly focusing on the perpetrator, rather than addressing the flaws in their platforms that allow them to be weaponized.

In their defense, Twitter, Facebook, Google, and the other big tech companies are in an incredibly difficult position. We’re asking them to do the public service of protecting our democratic discourse — a function normally reserved for government, that isn’t realistic even for very powerful companies. And even if tech executives genuinely want to cultivate a healthier public debate, their first responsibility is to run a healthy business that produces a financial return for their shareholders. The problem is that what’s best for business doesn’t align with the public interest.

So the question is: if we can’t trust Facebook and Twitter to police themselves, who else is going to do it?