1
United States: Playing Whack-a-Troll
In early July 2017, I was in Ukraine, nearing the end of my Fulbright grant. During the first six months of Donald Trumpâs term as president, I watched from another continent as the political culture of the United States grew increasingly charged. For the first time in my life, a majority of my friends and family were regularly attending protests and more passionately following politics. And it seemed that because of Trump and scrutiny over his connections to Putin, for the first time in my career, Americans suddenly had an increased interest in Russia and Eastern Europe.
So it wasnât a surpriseâmore of a curiosity of the new era, I supposeâthat a highly niche group in my orbit was protesting and politicking, too. While scientists I knew marched in opposition to the administrationâs assault on reason and my lawyer friends camped out at airports to offer help to visitors affected by the new travel ban on people from predominantly Muslim countries, the theater people in my life were getting ready to raise their voices against the administration as well. But true to their roots, they did it in song.
Iâve done theaterâmostly musicalsâsince I was a kid. When I moved to Washington, I was surprised to find a robust community theater scene in the area. It turned out the high-powered, obsessive, hardworking group that made the nationâs capital run could also put on a damn good show. The leading lady might be a top aide to a senator. In a production of Chess, a musical about the game as a Cold War proxy battle, a Republican opposition researcher played the American political operator trying to secure a Russian chess grandmasterâs defection. (The similarities between their characters were lost on no one, including the actor.) Sadly, we produced the show in 2015, a little over a year before caring about Russia was back in style.
After moving to Ukraine, I was on a forced hiatus from the DC theater scene for nearly two years. I was still friends with many of my former castmates and followed the local theater chat on Facebook, though, ever eager to live vicariously through those still onstage back at home. In early summer 2017, an event appeared in my feed for a protest outside of the White House on July 4, inviting âresistance activists, show tune lovers, and karaoke fans to ⌠sing a song of freedom and demand Trumpâs impeachment.â1 Hundreds of people were planning to dress in colonial attire at the height of Washingtonâs muggy summer to sing a parody of âDo You Hear the People Sing?,â the famous revolutionary anthem from the musical Les Miserables, in front of the so-called peopleâs house.
I chuckled and rolled my eyes at the idea of such a spectacle. âEverythingâs a performance opportunity for theater people,â I thought. The protest didnât cross my mind again for more than a year, until I read about it in a criminal complaint against the Internet Research Agency (IRA), Agency, the infamous troll factory in St. Petersburg, Russia.
*
In an entirely unexpected collision of my two great loves, it seemed that Russia had weaponized show tunes.
At the conclusion of the US investigation into Russian interference into the 2016 US presidential election, Special Counsel Robert Mueller closed his one and only public press appearance by âreiterating the central allegation of [the investigationâs] indictments: that there were multiple, systematic efforts to interfere in our election, and that allegation deserves the attention of every American.â2
Since revelations of Russiaâs on- and offline influence campaign came to light, and even since Muellerâs May 2019 warning, the United States still seems not to grasp the complexity of what has befallen our country. When newscasters, commentators, and politicians describe Russiaâs interference, they rely on a wide swath of vagariesâfake news, bots, trollsâthat they not only often fail to use correctly, but also fail to accurately describe the depth and breadth of Russiaâs interference in the United States and other countries before, during, and after the 2016 election. The problem is presented as either a curiosity of technology or the unforeseen capriciousness of an adversary, but in reality, neither more democratic social media nor better foreign policy forethought would have solved it entirely.
What makes this information war so difficult to win is not just the online tools that amplify and target its messages or the adversary that is sending them; itâs the fact that those messages are often unwittingly delivered not by trolls or bots, but by authentic local voices. When taken together with an understanding of the way Russia has operated for more than a decade in Eastern Europe, the Mueller Report makes it clear that this challenge is far greater than winning a game of Whack-a-Troll.
Russiaâs information warfare offensive against the United States began as early as 2014, intensified after the 2016 election, and continues to this day. It involves many layers and components and is more intricate than the purchase of $100,000 in Facebook ads or the simple activation of a few lines of computer code and a network of fake Twitter accounts to make topics trend. Facebookâs Mark Zuckerberg was not entirely wrong when he said âthe idea that fake news on Facebook ⌠influenced the election in any wayâI think itâs a pretty crazy idea.â Zuckerberg goes too far, of course, by saying that Facebook did not influence the election in any way; Facebook and Facebook-owned properties such as Instagram were the main vectors that Russia used to influence the discourse surrounding the election. But social media relies on user engagement with content to function, and Americans were quite happy to interact with and share sensationalist, divisive, and unfounded information they saw during the course of the election and beyond. What they didnât know is that they were sharing content produced and posted from accounts run out of Russiaâs Internet Research Agency.
Like the Eastern European Kremlin-sponsored influence operations that came before them, Russian online interference surrounding the 2016 US presidential election had the goal of âprovok[ing] and amplify[ing] political and social discord in the United States.â3 Durin g the 2016 US presidential election, the operation had the more explicit goal of denigrating Hillary Clinton and promoting Donald Trump. Whether or not a single vote changed as a result of Internet Research Agency operations, the IRAâs use of several interdependent components of online influence campaigns altered the content and tone of discourse during the election and beyond, in addition to changing day-to-day behavior; IRA campaigns got people across the political spectrum to show up to protest actions all over the country. This fact is not repeated nearly often enough, so I will restate it: by creating and tending to trusted communities of hundreds of thousands of individuals over the course of several years, Russian operatives were able to encourage some Americans to show up to protests. They changed behavior.
The IRAâs Facebook operation began with the cultivation of fake social media profiles in 2015, having monitored American social media groups âdedicated to US Politics and social issuesâ for the previous year.4 The monitoring had an in-person component as well; two IRA employees, Aleksandra Krylova and Anna Bogacheva, obtained visas on false pretenses and, on a whirlwind tour that would later get them indicted for conspiracy to defraud the United States, traveled to America for three weeks in 2014, visiting Nevada, New Mexico, Colorado, Illinois, Michigan, Texas, and New York.5 Another IRA employee traveled to Atlanta. Later, the IRA put these intelligence-gathering operations, as well as the information it had gleaned from monitoring social media pages and groups, to good use. Russian âspecialistsâ at the Internet Research Agency posing as American activists created and curated their own inauthentic groups and pages around social issues, including some targeting âanti-immigration groups, Tea Party activists, [and] Black Lives Matter protestors.â6 By the middle of 2016, the IRAâs unit targeting the US election had a staff of at least eighty individuals and a monthly budget of over $1,250,000.7
Far from simply regurgitating links to Russian propaganda outlets or outright fake stories, at first these groups sought to capitalize on positive emotion and build community. Two groupsâBlacktivist, which had more followers than the official Black Lives Matter Facebook page, and Being Patriotic, a right-wing page that channeled the spirit of Americana to the point of jingoismâfrequently employed this tactic. While building community on the Blacktivist page, the IRA regularly shared articles and memes that telegraphed pride in African American history and contributions to society. Being Patriotic shared pro-gun, pro-America content that was fairly inane, including my personal favorite: a picture of a smiling Golden Retriever wearing a red bandanna with white stars and âholdingâ an American flag. The text read, âLike if you think itâs going to be a great week!â Who would think twice before liking that patriotic dog or sharing their pride in their history with all of their Facebook friends? There seemed to be nothing malicious about it, but this was how the IRA built trust and cohesion on its social properties.
The IRA used the same tactics to target the LGBT community, American Muslims, and other groups situated around the fissures running through the US sociopolitical landscape. Race, religion, sexual orientation, and other sensitive areas that capitalize on distrust in government or the proverbial American âother sideâ were lurking behind a facade of positivity and community. Through this generally apolitical content, the Internet Research Agency built usersâ trust in the pages, increased engagement, and grew their followings to hundreds of thousands of people each.
On other social media platforms, including Instagram and Twitter, the IRA attracted five- and six-digit follower counts. One popular IRA-run Twitter account claimed to represent the Tennessee Republican Party (@TEN_GOP, and later, after its original account was deactivated, @10_GOP), while other accounts in the IRAâs portfolio represented themselves as Republican anti-immigration activists. These fake accounts gradually gained notoriety, earning media mentions in legitimate and well-respected outlets, as well as replies and retweets from high-powered users, including Kellyanne Conway and the Trump family. Then candidate Donald Trump himself replied to a tweet from the fake Tennessee Republican Party account that had highlighted a Trump event in Florida: âTHANK YOU for your support, Miami! My team just shared photos from your TRUMP SIGN WAVING DAY, yesterday! I love youâand there is no questionâTOGETHER, WE WILL MAKE AMERICA GREAT AGAIN!â8 The tweet has since been deleted.
Whether on Facebook, Twitter, Instagram, or beyond, these communities were built on emotion, not through the purchase of online advertising. Not only had IRA operatives created fake profiles and fake groups, they had duped real, sometimes influential, Americans into interacting with them.
The operation did not end with enticing patriotic dog pictures. IRA employees had been instructed to instigate âpolitical intensityâ by âsupporting radical groups, users dissatisfied with [the] social and economic situations and oppositional social movements.â9 As trust within their inauthentic communities grew, the IRA operativesâ asks of community members did as well. They began by initiating armchair activist campaigns based on easy tasks that furthered the sense of community in the group. The very first Facebook advertisement the IRA purchased âexplicitly endorsing the Trump campaignâ encouraged followers of the âTea Party Newsâ Instagram account to âmake a patriotic team of young Trump supporters by uploading photos with the hashtag #KIDS4TRUMP.â10 Another ad on âBlack Matters,â a Facebook page targeting African Americans, encouraged followers to âWear black, fight back!â as part of a national day of protest against âpolice brutality, repression, and criminalization of a generation.â It solicited pictures from people who attended the protest event and promised to post them on the page âso that the whole African American community will know about you and your efforts.â Across other pages, IRA specialists implored users to change their profile pictures in support of a cause or sign petitions about political issues.
The IRA also purchased 3,500 advertisements on Facebook to boost the engagement and reach on their posts. According to information that Facebook disclosed, the advertising campaign cost about $100,000. Journalists and politicians who focused American discourse on this part of the operation were shocked to find that Russians could use rubles to purchase an ad that would influence the American election the same way cigarette company might place an ad to target vulnerable potential buyers. While this concern does bring up valid question s about the regulation of online advertisements in election periods, itâs misplaced with regard to Russia; the operation was already successful without them.
This became clear in May 2018, when Democrats from the House of Representativesâ Permanent Select Committee on Intelligence released a library of the IRAâs ad purchases. It was the first largely unredacted public look at the type of content the Internet Research Agency had been creating and amplifying. Some of the posts were riddled with English mistakes. (âBlack lives matters!â read the text of one ad.) Others were just odd; the IRA advertised an Instagram account called âLiberty Risingâ which targeted followers of the popular online comedy outlet College Humor. According to an ad, Liberty Rising claimed to share âall the funniest memes and actual news. Read us or die suffering.â Several of the ads, which were released with their engagement metrics, performed extremely poorly. As a result, some journalists and commentators brushed off the IRAâs ad purchases as unsophisticated and ineffective. If these bumbling efforts were the extent of their interference, how could they possibly have changed any Americansâ votes, they asked. But this wasnât the full extent of the Russian online influence campaign; contrarians chuckling at the IRAâs alleged ineffectiveness ignored that the promoted posts performed well on their own, completely organically, without help from ads, thanks to the engaged communities that the IRA built over a period of several years. A post that might have only been clicked a few times by those to whom the ad was served would boast hundreds or sometimes thousands of engagements from the groups where the content was initially shared.
Researchers at the Oxford Internet Institute confirmed this trend. With access to the House-released advertisements and another dataset of accounts of IRA origin that the Senate Select Committee on Intelligence provided them, the researchers found that of 67,502 organic Facebook posts across eighty-one different pages, and of 116,205 organic Instagram posts across 133 different accounts, organic contentânot advertisementsâdrove the bulk of IRA activity and engagement on Facebook-owned platforms.11 The organic Facebook posts in the dataset âwere shared by users just under 31 million times, liked almost 39 million times, reacted to with emojis almost 5.4 billion times, and ⌠generat[ed] almost 3.5 million comments.â
All of these details are more or less accepted today. But a detail that deserves more attention, and one that scares me, is that the IRA operation also had an offlineâor as internet users sometimes say, an âIRLâ (in real life)âcomponent. It combined online advertising with the IRAâs intimate knowledge of the fissures crisscrossing the American political landscape to turn armchair activism into real rallies on the streets of the United States.
*
On October 19, 2018, a criminal complaint in Special Counsel Robert Muellerâs investigation into Russian interference in the 2016 election was unsealed. It lays out how ...