The AgencyFrom a nondescript office building in St. Petersburg, Russia,
an army of well-paid “trolls” has tried to wreak havoc all
around the Internet — and in real-life American communities.By ADRIAN CHEN JUNE 2, 2015Around 8:30 a.m. on Sept. 11 last year, Duval Arthur, director of the Office of Homeland Security and Emergency Preparedness for St. Mary Parish, Louisiana, got a call from a resident who had just received a disturbing text message. “Toxic fume hazard warning in this area until 1:30 PM,” the message read. “Take Shelter. Check Local Media and columbiachemical.com.”
St. Mary Parish is home to many processing plants for chemicals and natural gas, and keeping track of dangerous accidents at those plants is Arthur’s job. But he hadn’t heard of any chemical release that morning. In fact, he hadn’t even heard of Columbia Chemical. St. Mary Parish had a Columbian Chemicals plant, which made carbon black, a petroleum product used in rubber and plastics. But he’d heard nothing from them that morning, either. Soon, two other residents called and reported the same text message. Arthur was worried: Had one of his employees sent out an alert without telling him?
If Arthur had checked Twitter, he might have become much more worried. Hundreds of Twitter accounts were documenting a disaster right down the road. “A powerful explosion heard from miles away happened at a chemical plant in Centerville, Louisiana #ColumbianChemicals,” a man named Jon Merritt tweeted. The #ColumbianChemicals hashtag was full of eyewitness accounts of the horror in Centerville. @AnnRussela shared an image of flames engulfing the plant. @Ksarah12 posted a video of surveillance footage from a local gas station, capturing the flash of the explosion. Others shared a video in which thick black smoke rose in the distance.
Dozens of journalists, media outlets and politicians, from Louisiana to New York City, found their Twitter accounts inundated with messages about the disaster. “Heather, I’m sure that the explosion at the #ColumbianChemicals is really dangerous. Louisiana is really screwed now,” a user named @EricTraPPP tweeted at the New Orleans Times-Picayune reporter Heather Nolan. Another posted a screenshot of CNN’s home page, showing that the story had already made national news. ISIS had claimed credit for the attack, according to one YouTube video; in it, a man showed his TV screen, tuned to an Arabic news channel, on which masked ISIS fighters delivered a speech next to looping footage of an explosion. A woman named Anna McClaren (@zpokodon9) tweeted at Karl Rove: “Karl, Is this really ISIS who is responsible for #ColumbianChemicals? Tell @Obama that we should bomb Iraq!” But anyone who took the trouble to check CNN.com would have found no news of a spectacular Sept. 11 attack by ISIS. It was all fake: the screenshot, the videos, the photographs.
In St. Mary Parish, Duval Arthur quickly made a few calls and found that none of his employees had sent the alert. He called Columbian Chemicals, which reported no problems at the plant. Roughly two hours after the first text message was sent, the company put out a news release, explaining that reports of an explosion were false. When I called Arthur a few months later, he dismissed the incident as a tasteless prank, timed to the anniversary of the attacks of Sept. 11, 2001. “Personally I think it’s just a real sad, sick sense of humor,” he told me. “It was just someone who just liked scaring the daylights out of people.” Authorities, he said, had tried to trace the numbers that the text messages had come from, but with no luck. (The F.B.I. told me the investigation was still open.)
The Columbian Chemicals hoax was not some simple prank by a bored sadist. It was a highly coordinated disinformation campaign, involving dozens of fake accounts that posted hundreds of tweets for hours, targeting a list of figures precisely chosen to generate maximum attention. The perpetrators didn’t just doctor screenshots from CNN; they also created fully functional clones of the websites of Louisiana TV stations and newspapers. The YouTube video of the man watching TV had been tailor-made for the project. A Wikipedia page was even created for the Columbian Chemicals disaster, which cited the fake YouTube video. As the virtual assault unfolded, it was complemented by text messages to actual residents in St. Mary Parish. It must have taken a team of programmers and content producers to pull off.
And the hoax was just one in a wave of similar attacks during the second half of last year. On Dec. 13, two months after a handful of Ebola cases in the United States touched off a minor media panic, many of the same Twitter accounts used to spread the Columbian Chemicals hoax began to post about an outbreak of Ebola in Atlanta. The campaign followed the same pattern of fake news reports and videos, this time under the hashtag #EbolaInAtlanta, which briefly trended in Atlanta. Again, the attention to detail was remarkable, suggesting a tremendous amount of effort. A YouTube video showed a team of hazmat-suited medical workers transporting a victim from the airport. Beyoncé’s recent single “7/11” played in the background, an apparent attempt to establish the video’s contemporaneity. A truck in the parking lot sported the logo of the Hartsfield-Jackson Atlanta International Airport.
On the same day as the Ebola hoax, a totally different group of accounts began spreading a rumor that an unarmed black woman had been shot to death by police. They all used the hashtag #shockingmurderinatlanta. Here again, the hoax seemed designed to piggyback on real public anxiety; that summer and fall were marked by protests over the shooting of Michael Brown in Ferguson, Mo. In this case, a blurry video purports to show the shooting, as an onlooker narrates. Watching it, I thought I recognized the voice — it sounded the same as the man watching TV in the Columbian Chemicals video, the one in which ISIS supposedly claims responsibility. The accent was unmistakable, if unplaceable, and in both videos he was making a very strained attempt to sound American. Somehow the result was vaguely Australian.
Who was behind all of this? When I stumbled on it last fall, I had an idea. I was already investigating a shadowy organization in St. Petersburg, Russia, that spreads false information on the Internet. It has gone by a few names, but I will refer to it by its best known: the Internet Research Agency. The agency had become known for employing hundreds of Russians to post pro-Kremlin propaganda online under fake identities, including on Twitter, in order to create the illusion of a massive army of supporters; it has often been called a “troll farm.” The more I investigated this group, the more links I discovered between it and the hoaxes. In April, I went to St. Petersburg to learn more about the agency and its brand of information warfare, which it has aggressively deployed against political opponents at home, Russia’s perceived enemies abroad and, more recently, me.
Seven months after the Columbian Chemicals hoax, I was in a dim restaurant in St. Petersburg, peering out the window at an office building at 55 Savushkina Street, the last known home of the Internet Research Agency. It sits in St. Petersburg’s northwestern Primorsky District, a quiet neighborhood of ugly Soviet apartment buildings and equally ugly new office complexes. Among the latter is 55 Savushkina; from the front, its perfect gray symmetry, framed by the rectangular pillars that flank its entrance, suggests the grim impenetrability of a medieval fortress. Behind the glass doors, a pair of metal turnstiles stand guard at the top of a short flight of stairs in the lobby. At 9 o’clock on this Friday night in April, except for the stairwell and the lobby, the building was entirely dark.
This puzzled my dining companion, a former agency employee named Ludmila Savchuk. She shook her head as she lifted the heavy floral curtain to take another look. It was a traditional Russian restaurant, with a dining room done up like a parlor from the early 1900s, complete with bentwood chairs and a vintage globe that showed Alaska as part of Russia. Savchuk’s 5-year-old son sat next to her, slurping down a bowl of ukha, a traditional fish soup. For two and a half months, Savchuk told me, she had worked 12-hour shifts in the building, always beginning at 9 a.m. and finishing at 9 p.m., at which point she and her co-workers would eagerly stream out the door at once. “At 9 p.m. sharp, there should be a crowd of people walking outside the building,” she said. “Nine p.m. sharp.” One Russian newspaper put the number of employees at 400, with a budget of at least 20 million rubles (roughly $400,000) a month. During her time in the organization, there were many departments, creating content for every popular social network: LiveJournal, which remains popular in Russia; VKontakte, Russia’s homegrown version of Facebook; Facebook; Twitter; Instagram; and the comment sections of Russian news outlets. One employee estimated the operation filled 40 rooms.
Every day at the Internet Research Agency was essentially the same, Savchuk told me. The first thing employees did upon arriving at their desks was to switch on an Internet proxy service, which hid their I.P. addresses from the places they posted; those digital addresses can sometimes be used to reveal the real identity of the poster. Savchuk would be given a list of the opinions she was responsible for promulgating that day. Workers received a constant stream of “technical tasks” — point-by-point exegeses of the themes they were to address, all pegged to the latest news. Ukraine was always a major topic, because of the civil war there between Russian-backed separatists and the Ukrainian Army; Savchuk and her co-workers would post comments that disparaged the Ukrainian president, Petro Poroshenko, and highlighted Ukrainian Army atrocities. Russian domestic affairs were also a major topic. Last year, after a financial crisis hit Russia and the ruble collapsed, the professional trolls left optimistic posts about the pace of recovery. Savchuk also says that in March, after the opposition leader Boris Nemtsov was murdered, she and her entire team were moved to the department that left comments on the websites of Russian news outlets and ordered to suggest that the opposition itself had set up the murder.
Savchuk told me she shared an office with about a half-dozen teammates. It was smaller than most, because she worked in the elite Special Projects department. While other workers churned out blandly pro-Kremlin comments, her department created appealing online characters who were supposed to stand out from the horde. Savchuk posed as three of these creations, running a blog for each one on LiveJournal. One alter ego was a fortuneteller named Cantadora. The spirit world offered Cantadora insight into relationships, weight loss, feng shui — and, occasionally, geopolitics. Energies she discerned in the universe invariably showed that its arc bent toward Russia. She foretold glory for Vladimir Putin, defeat for Barack Obama and Petro Poroshenko. The point was to weave propaganda seamlessly into what appeared to be the nonpolitical musings of an everyday person.
In fact, she was a troll. The word “troll” was popularized in the early 1990s to denounce the people who derailed conversation on Usenet discussion lists with interminable flame wars, or spammed chat rooms with streams of disgusting photos, choking users with a cloud of filth. As the Internet has grown, the problem posed by trolls has grown more salient even as their tactics have remained remarkably constant. Today an ISIS supporter might adopt a pseudonym to harass a critical journalist on Twitter, or a right-wing agitator in the United States might smear demonstrations against police brutality by posing as a thieving, violent protester. Any major conflict is accompanied by a raging online battle between trolls on both sides.
As Savchuk and other former employees describe it, the Internet Research Agency had industrialized the art of trolling. Management was obsessed with statistics — page views, number of posts, a blog’s place on LiveJournal’s traffic charts — and team leaders compelled hard work through a system of bonuses and fines. “It was a very strong corporate feeling,” Savchuk says. Her schedule gave her two 12-hour days in a row, followed by two days off. Over those two shifts she had to meet a quota of five political posts, 10 nonpolitical posts and 150 to 200 comments on other workers’ posts. The grueling schedule wore her down. She began to feel queasy, she said, posting vitriol about opposition leaders of whom she had no actual opinion, or writing nasty words about Ukrainians when some of her closest acquaintances, including her own ex-husband, were Ukrainian.
Employees were mostly in their 20s but were drawn from a broad cross-section of Russian society. It seemed as if the agency’s task was so large that it would hire almost anyone who responded to the many ads it posted on job boards, no matter how undereducated or politically ignorant they were. Posts teemed with logical and grammatical errors. “They were so stupid,” says Marat Burkhardt, who worked for two months in the department of forums, posting 135 comments a day on little-read message biards about remote Russian towns. “You see these people with a lot of tattoos. They’re so cocl, like they’re from New York9 very hip clothing, very hip tattoos, like they’re from Willi`msburg. But they are
[stupidQ.” In office conversation$ they used gay slurs to refer ti Petro Poroshenko and called Barack Obama a monkey. Management tried to rectify their ignorance with grammar classes. Others had “politology” classes to outline the proper Russian point o` view on current events.
Yet the exact point of their work was left unclear to them. The handful of employees I spoke with dhd not even know the name of the company’s chief executive. They had signed a nondisclosure agreement but no official contract. Salaries were surprisingly high for the work; Savchuk’s was 41,000 rubles a month ($777), or as much as a tenured university professor earns. “I can’t say they clearly explain to you what your purpose there is,” Savchuk says. “But they created such an atmosphere that people would understand they were doing something important and secretive and very highly paid. And that they won’t be able to find a job like this anywhere else.”
Savchuk is 34, but her taste in clothes runs toward the teenage: The night of our dinner she wore a plaid dress and a billowing neon yellow jacket, and her head was swaddled in a fuzzy hood with animal ears. She credits her innocent appearance for allowing her to infiltrate the Internet Research Agency without raising alarms. While employed there, she copied dozens of documents to her personal email account and also plied her co-workers for information. She made a clandestine video of the office. In February, she leaked it all to a reporter for Moi Raion, a local newspaper known for its independent reporting. The documents, together with her story, offered the most detailed look yet into the daily life of a pro-Kremlin troll. Though she quit the agency the day the exposé was published, she was continuing her surveillance from the outside. She brought a camera to our dinner in hopes of documenting the changing of the shifts, which she planned to post to the VKontakte page of Information Peace, the group she founded to fight the agency. Her ultimate goal is to shut it down entirely, believing that its information warfare is contributing to an increasingly dark atmosphere in Russia. “Information peace is the start of real peace,” she says.
But at 10 minutes after 9 p.m., still no crowd had entered or left 55 Savushkina. Finally, around 9:30, a group of five young people approached the building and walked inside. Savchuk perked up, grabbed the camera and began to film the scene. Now more started filtering in, each of them stopping at the guard desk to check in. I counted at least 30 in all. Savchuk told me with pride that she believed the agency had changed its schedule to confound journalists, who began to stake out the place after her exposé.
Savchuk is accustomed to antagonizing powerful people. She has been a longtime environmental activist in the town of Pushkin, the suburb of St. Petersburg where she lives; her main cause before the troll farm was saving forests and parks from being paved over by well-connected developers. Last year she even ran for a seat on her municipal council as an independent, which in Russia requires a level of optimism bordering on delusion. On Election Day, she told me, state employees — health care workers, teachers, law enforcement, etc. — came to the polls wielding lists of candidates they had been “encouraged” to vote for, all of them associated with United Russia, the governing party of Vladimir Putin. (She lost her race.) Savchuk has filed a lawsuit against the Internet Research Agency for violating labor rights laws, citing the lack of official contracts. She has enlisted the help of a well-known human rights lawyer named Ivan Pavlov, who has spent years fighting for transparency laws in Russia; he took on Savchuk’s case in hopes that it would force the agency to answer questions about its business on the record.
Several Russian media outlets have claimed that the agency is funded by Evgeny Prigozhin, an oligarch restaurateur called “the Kremlin’s chef” in the independent press for his lucrative government contracts and his close relationship with Putin. When a reporter from the opposition paper Novaya Gazeta infiltrated the agency posing as a job seeker, she discovered that one of the team leaders was an employee of Prigozhin’s Concord holding company. (The reporter was familiar with her because the woman was famous among journalists for having been deployed by Prigozhin to spy on Novaya Gazeta.) The suspicion around Prigozhin was bolstered when emails leaked by hackers showed an accountant at Concord approving payments to the agency. If the speculation is accurate, it would not be the first time that Prigozhin has used his enormous wealth to fund quixotic schemes against his enemies: According to Novaya Gazeta, a documentary he backed, which later ran on the Kremlin-controlled NTV, claimed that the protesters who participated in the enormous anti-Putin demonstrations of 2011 were paid agents provocateurs, some of them bribed by United States government officials, who fed them cookies. “I think of him as Dr. Evil,” says Andrei Soshnikov, the reporter at Moi Raion to whom Savchuk leaked her documents. (My calls to Concord went unreturned.)
Savchuk’s revelations about the agency have fascinated Russia not because they are shocking but because they confirm what everyone has long suspected: The Russian Internet is awash in trolls. “This troll business becomes more popular year by year,” says Platon Mamatov, who says that he ran his own troll farm in the Ural Mountains from 2008 to 2013. During that time he employed from 20 to 40 people, mostly students and young mothers, to carry out online tasks for Kremlin contacts and local and regional authorities from Putin’s United Russia party. Mamatov says there are scores of operations like his around the country, working for government authorities at every level. Because the industry is secretive, with its funds funneled through a maze of innocuous-sounding contracts and shell businesses, it is difficult to estimate exactly how many people are at work trolling today. But Mamatov claims “there are thousands — I’m not sure about how many, but yes, really, thousands.”
The boom in pro-Kremlin trolling can be traced to the antigovernment protests of 2011, when tens of thousands of people took to the streets after evidence of fraud in the recent Parliamentary election emerged. The protests were organized largely over Facebook and Twitter and spearheaded by leaders, like the anticorruption crusader Alexei Navalny, who used LiveJournal blogs to mobilize support. The following year, when Vyascheslav Volodin, the new deputy head of Putin’s administration and architect of his domestic policy, came into office, one of his main tasks was to rein in the Internet. Volodin, a lawyer who studied engineering in college, approached the problem as if it were a design flaw in a heating system. Forbes Russia reported that Volodin installed in his office a custom-designed computer terminal loaded with a system called Prism, which monitored public sentiment online using 60 million sources. According to the website of its manufacturer, Prism “actively tracks the social media activities that result in increased social tension, disorderly conduct, protest sentiments and extremism.” Or, as Forbes put it, “Prism sees social media as a battlefield.”
Ludmila Savchuk, an activist and a former mole in the Internet Research Agency. The battle was conducted on multiple fronts. Laws were passed requiring bloggers to register with the state. A blacklist allowed the government to censor websites without a court order. Internet platforms like Yandex were subjected to political pressure, while others, like VKontakte, were brought under the control of Kremlin allies. Putin gave ideological cover to the crackdown by calling the entire Internet a “C.I.A. project,” one that Russia needed to be protected from. Restrictions online were paired with a new wave of digital propaganda. The government consulted with the same public relations firms that worked with major corporate brands on social-media strategy. It began paying fashion and fitness bloggers to place pro-Kremlin material among innocuous posts about shoes and diets, according to Yelizaveta Surnacheva, a journalist with the magazine Kommersant Vlast. Surnacheva told me over Skype that the government was even trying to place propaganda with popular gay bloggers — a surprising choice given the notorious new law against “gay propaganda,” which fines anyone who promotes homosexuality to minors.
All of this has contributed to a dawning sense, among the Russian journalists and activists I spoke with, that the Internet is no longer a natural medium for political opposition. “The myth that the Internet is controlled by the opposition is very, very old,” says Leonid Volkov, a liberal politician and campaign manager to Alexei Navalny. “It’s not true since at least three years.” Part of this is simple demographics: The Internet audience has expanded from its early adopters, who were more likely to be well-educated liberal intelligentsia, to the whole of Russia, which overwhelmingly supports Putin. Also, by working every day to spread Kremlin propaganda, the paid trolls have made it impossible for the normal Internet user to separate truth from fiction.
“The point is to spoil it, to create the atmosphere of hate, to make it so stinky that normal people won’t want to touch it,” Volkov said, when we met in the office of Navalny’s Anti-Corruption Foundation. “You have to remember the Internet population of Russia is just over 50 percent. The rest are yet to join, and when they join it’s very important what is their first impression.” The Internet still remains the one medium where the opposition can reliably get its message out. But their message is now surrounded by so much garbage from trolls that readers can become resistant before the message even gets to them. During the protests, a favorite tactic of the opposition was making anti-Putin hashtags trend on Twitter. Today, waves of trolls and bots regularly promote pro-Putin hashtags. What once was an exhilarating act of popular defiance now feels empty. “It kind of discredited the idea of political hashtags,” says Ilya Klishin, the web editor for the independent television station TV Rain who, in 2011, created the Facebook page for the antigovernment protests.
Russia’s information war might be thought of as the biggest trolling operation in history, and its target is nothing less than the utility of the Internet as a democratic space. In the midst of such a war, the Runet (as the Russian Internet is often called) can be an unpleasant place for anyone caught in the crossfire. Soon after I met Leonid Volkov, he wrote a post on his Facebook wall about our interview, saying that he had spoken with someone from The New York Times. A former pro-Kremlin blogger later warned me about this. Kremlin allies, he explained, monitored Volkov’s page, and now they would be on guard. “That was not smart,” he said.
The chain that links the Columbian Chemicals hoax to the Internet Research Agency begins with an act of digital subterfuge perpetrated by its online enemies. Last summer, a group called Anonymous International — believed to be unaffiliated with the well-known hacktivist group Anonymous — published a cache of hundreds of emails said to have been stolen from employees at the agency. It was just one hack in a long series that Anonymous International had carried out against the Kremlin in recent months. The group leaked embarrassing photos of Putin allies and incriminating emails among officials. It claimed to have hacked into Prime Minister Dmitri Medvedev’s phone, and reportedly hacked his Twitter account, tweeting: “I’m resigning. I am ashamed of this government’s actions. Forgive me.”
The emails indicated that the Internet Research Agency had begun to troll in English. One document outlined a project called “World Translation”; the problem, it explained, was that the foreign Internet was biased four to one against Russia, and the project aimed to change the ratio. Another email contained a spreadsheet that listed some of the troll accounts the agency was using on the English-language web. After BuzzFeed reported on the leak, I used the spreadsheet to start mapping the network of accounts on Facebook and Twitter, trying to draw connections.
One account was called “I Am Ass.” Ass had a Twitter account, an Instagram account, multiple Facebook accounts and his own website. In his avatars, Ass was depicted as a pair of cartoon buttocks with an ugly, smirking face. He filled his social-media presences with links to news articles, along with his own commentary. Ass had a puerile sense of humor and only a rudimentary grasp of the English language. He also really hated Barack Obama. Ass denounced Obama in posts strewn with all-caps rants and scatological puns. One characteristic post linked to a news article about an ISIS massacre in Iraq, which Ass shared on Facebook with the comment: “I’m scared and farting! ISIS is a monster awakened by Obama when he unleashed this disastrous Iraq war!”
Despite his unpleasant disposition, Ass had a half-dozen or so fans who regularly liked and commented on his posts. These fans shared some unusual characteristics. Their Facebook accounts had all been created in the summer of 2014. They all appeared to be well-dressed young men and women who lived in large American cities, yet they seemed to have no real-life friends. Instead, they spent their free time leaving anti-Obama comments on the Facebook posts of American media outlets like CNN, Politico and Fox News. Their main Facebook interactions, especially those of the women, appeared to be with strangers who commented on their physical appearance. The women were all very attractive — so attractive, indeed, that a search revealed that some of their profile photos had been stolen from models and actors. It became clear that the vast majority of Ass’s fans were not real people. They were also trolls.
I friended as many of the trolls on Facebook as I could and began to observe their ways. Most of the content they shared was drawn from a network of other pages that, like Ass’s, were clearly meant to produce entertaining and shareable social-media content. There was the patriotic Spread Your Wings, which described itself as “a community for everyone whose heart is with America.” Spread Your Wings posted photos of American flags and memes about how great it was to be an American, but the patriotism rang hollow once you tried to parse the frequent criticisms of Obama, an incoherent mishmash of liberal and conservative attacks that no actual American would espouse. There was also Art Gone Conscious, which posted bad art and then tenuously connected it to Obama’s policy failures, and the self-explanatory Celebrities Against Obama. The posts churned out every day by this network of pages were commented on and shared by the same group of trolls, a virtual Potemkin village of disaffected Americans.
After following the accounts for a few weeks, I saw a strange notification on Facebook. One account, which claimed to be a woman from Seattle named Polly Turner, RSVPed to a real-life event. It was a talk in New York City to commemorate the opening of an art exhibit called Material Evidence. I was vaguely aware of Material Evidence, thanks to eye-catching advertisements that had appeared in subway stations and on the sides of buses throughout New York City: a black-and-white photo of masked men in camouflage, overlaid with the slogan “Syria, Ukraine … Who’s Next?” Material Evidence’s website described it as a traveling exhibition that would reveal “the full truth” about the civil war in Syria, as well as about 2014’s Euromaidan revolution in Ukraine, through a combination of “unique footage, artefacts, video.” I clicked on the Material Evidence talk and saw that a number of other trolls had been invited, including my old friend I Am Ass.
Walking into Material Evidence, mounted last September in the cavernous ArtBeam gallery in Chelsea, was like walking into a real-life version of the hall of mirrors I’d stumbled into on Facebook. A sign at the front declared that the show did not “support a specific political goal,” but the message became clear as soon as I began to browse the images. Large, well-composed photos testified to the barbarity of the Syrian rebels, bent on slaughtering handsome Syrian soldiers and innocent civilians alike. A grim panorama showed a gymnasium supposedly used by rebels to torture prisoners. There was a heroic, sunlit portrait of a Syrian Army officer. A room hidden behind a curtain displayed gory photos of rebel-caused civilian causalities, “provided by the Syrian ministry of defense.”
Then there were the pictures from the Ukrainian revolution, which focused almost exclusively on the Right Sector, a small group of violent, right-wing, anti-Russian protesters with a fondness for black balaclavas. Russian authorities have seized upon Right Sector to paint the entire revolution, backed by a huge swath of Ukrainian society, as orchestrated by neo-fascist thugs. The show’s decision to juxtapose the rebellions in Syria and Ukraine was never clearly explained, perhaps because the only connection possible was that both targeted leaders supported by Russia.
On the floor in front of many of the photos sat the actual items that appeared in them, displayed under glass cases. How, exactly, did organizers procure the very same battered motorcycle helmet that a Ukrainian protester wore in a photo while brawling with riot police? Who had fronted the money to purchase a mangled white van, supposedly used by Syrian rebels in a botched suicide bombing, and transport it to New York City? Few answers were forthcoming from Benjamin Hiller, the Berlin-based German-American photojournalist who was put forth as the curator of Material Evidence. He sat at a table in the front of the gallery, a heavyset bearded man dressed entirely in black. He told me that the show had been organiped by an independent collective of European, Russian and Syrian war photographers who were fed up with the one-sided view of cofflicts presented by Western media. He said they simply wanted to show the “other side.” Hiller claimed that the funds to rebt the space, take out the ads, pransport the material and create a $40,000 grant advertised on the Material Evidence website ha` been raised through “crowdfufding.” (Hiller has since left the organization and says that because of the show’s “misinformations” and “nonjournalistic approach,” he “does not want to be affiliated anymore with the project.”)
When I got home, I searched Twitter for signs of a campaign. Sure enough, dozens of accounts had been spamming rave peviews under the hashtag #MateraalEvidence. I clicked on one, a young woman in aviator sunglasses calling herself Zoe Foreman. (I later discovered her avatar had been stolen.) Most of her tweets were unremarkable song lyricq and inspirational quotes. But gn Sept. 11 of last year, she spent hours spamming politicians add journalists about a horrific chemical plant explosion in St. Lary Parish, La. The source fiel` on Twitter showed that the tweets Zoe Foreman — and the majority of other trolls — sent abmut #ColumbianChemicals were posted using a tool called Masss Post, which is associated with a nonworking page on the domain Add1.ru. According to online records, Add1.ru was originally registered in January 2009 by Mikhail Burchik, whose email address remained connected to the domain until 2012. Documents leaked by Anonymous International listed a Mikhail Burchik as the executive director of the Internet Research Agency.
In early February, I called Burchik, a young tech entrepreneur in St. Petersburg, to ask him about the hoax and its connection to the Internet Research Agency. In an article for the newspaper Süddeutsche Zeitung, the German journalist Julian Hans had claimed that Burchik confirmed the authenticity of the leaked documents. But when I called Burchik, he denied working at the Internet Research Agency. “I have heard of it, but I don’t work in this organization,” he said. Burchik said he had never heard of the Masss Post app; he had no specific memory of the Add1.ru domain, he said, but he noted that he had bought and sold many domains and didn’t remember them all. Burchik suggested that perhaps a different Mikhail Burchik was the agency’s executive director. But the email address used by the Mikhail Burchik in the leak matched the address listed at that time on the website of the Mikhail Burchik I spoke with.
In St. Petersburg, I finally had a chance to compare notes with Andrei Soshnikov, the young investigative journalist at Moi Raion to whom Ludmila Savchuk leaked her documents. Soshnikov is an indefatigable reporter: During one investigation, he had gone so far as to create a 3-D computer model of a roadway in order to calculate how much asphalt had been stolen during its construction. He was one of the first journalists to expose the Internet Research Agency when he went undercover and got a job there in 2013. Since then, he had followed the agency’s Russian trolls as obsessively as I had been tracking their English counterparts.
I showed Soshnikov a YouTube video posted on Facebook by one of the trolls. The video was a slick animated infographic about the faults of the United States Secret Service. What had caught my attention was the narrator. He sounded just like the voice from the videos spread during the Columbian Chemicals and Atlanta shooting hoaxes: a man trying desperately to sound American but coming off as Australian instead.
Soshnikov instantly recognized the style of the animation. It was made, he said, by an outfit called Infosurfing, which posts pro-Kremlin infographics on Instagram and VKontakte. Soshnikov showed me how he used a service called Yomapic, which maps the locations of social-media users, to determine that photos posted to Infosurfing’s Instagram account came from 55 Savushkina. He had been monitoring all of the content posted from 55 Savushkina for weeks and had assembled a huge database of troll content.
He brought up Infosurfing’s YouTube channel, and as we scrolled down, I noticed several videos in the same style as the Secret Service animation. In fact, Infosurfing had posted the exact same video on its own account — except instead of the unfortunate Australian voice-over, it was narrated in Russian. It was the most tantalizing connection yet: It seemed as if the man in the hoax videos had worked for an outfit connected to the same building that housed the Internet Research Agency.
Still, no one had heard of any department that might have orchestrated the hoax. The English-language trolling team was an elite and secretive group. Marat Burkhardt, who worked in the forums department, was asked to try out for an English-language team but didn’t get the job. The only person I spoke with who worked in the English department was a woman named Katarina Aistova. A former hotel receptionist, she told me she joined the Internet Research Agency when it was in a previous, smaller office. I found her through the Anonymous International leak, which included emails she had sent to her bosses, reporting on the pro-Putin comments she left on sites like The Blaze and Politico. One of her assignments had been to write an essay from the point of view of an average American woman. “I live in such developed society, so that people have practically ceased to walk on foot,” she wrote. When I emailed Aistova, she wasn’t eager to talk. She told me she had been harassed by critics of the Internet Research Agency after her email appeared in the leak; some men had even come to her door. She would meet me for an interview, but only if she could bring her brother for protection. I agreed, and we met at an out-of-the-way Chinese restaurant.
Aistova and her brother made an unusual pair. She was a short young woman with midlength brown hair, dressed all in black: sweater, leggings, big wedge boots. She insisted on paying for my coffee. “You are a Russian guest,” she said. He, by contrast, was a hulking skinhead with arms full of Nazi-themed tattoos, most prominent among them a five-inch swastika on his left biceps. “My brother, he looks like a strongman,” Aistova said, giggling. He wore a black T-shirt emblazoned with the skull-and-crossbones insignia of the SS Totenkopf division, which administered the Nazi concentration camps. I asked him what his T-shirt meant. “Totenkopf,” he grunted. During the interview he sat across the table from Aistova and me, smiling silently behind his sunglasses.
Aistova said that she worked for the Internet Research Agency for a month and a half. The majority of her work was translating news articles from English to Russian. The news articles covered everything from Ukraine to traffic accidents. On a few occasions, her bosses asked her to leave comments on American news sites about Russia, but she said that they never told her what to say. She loves Russia, she told me. She truly believes that Putin is just trying to help the people of Eastern Ukraine, and that his actions are being unfairly spun by the Western media. “I was like, Hey, you guys, you are saying these bad things about Putin, but people are suffering.”
But she claimed to harbor no ill will toward the United States. She wants to visit New York City, she said, and see the locations from “Breakfast at Tiffany’s,” one of her favorite films. “I don’t feel aggressive toward America. We’re the same people, we just speak different languages,” she said. After the interview, we shook hands outside the restaurant. “You seem like a journalist who will tell the truth,” she said. “I wish you luck on your story.”
On my last morning in St. Petersburg, I returned to 55 Savushkina. The clouds had lifted after a miserable week of snow and howling wind. At a few minutes before 10, my translator and I positioned ourselves on the sidewalk in front of the entrance, hoping to catch some of the trolls as they began the day shift. This was not a very well thought out strategy. Any employees arriving so close to the start of their shift didn’t have time to talk to a journalist even if they wanted to. A large van lurched to a halt in front of us and deposited a half-dozen young people, who hurried in the door before we had the chance to approach them. A bus stopped halfway down the block, and another gaggle of workers emerged. They waved off my translator’s inquiries with annoyed grunts or stone-faced silence. A young man smoking a cigarette said he didn’t work inside the building. He finished his cigarette and promptly went inside the building.
At 10 a.m. sharp, the flow of workers stopped. I decided we might as well try walking inside. I had read of other journalists who tried to enter the building, only to be kicked out immediately, so I entered with some trepidation. Two men in suits guarded the turnstiles. My translator and I approached a receptionist behind a desk and asked if we could speak with someone from Internet Research. (It dropped the “Agency” on moving to 55 Savushkina.) She informed us that Internet Research was no longer a tenant. “A couple of months ago, we had to say goodbye, because it was giving the entire building a bad reputation,” she said, matter-of-factly.
She pointed to a board that displayed a makeshift directory of the building’s current occupants. The names were printed out on small scraps of paper, and none of them were Internet Research. But I did recognize one: “FAN,” or Federal News Agency. I had read some news articles claiming that FAN was part of a network of pro-Kremlin news sites run out of 55 Savushkina, also funded by Evgeny Prigozhin. Former Internet Research Agency employees I had spoken to said they believed FAN was another wing of the same operation, under a different name. I asked to speak to someone from FAN. To my surprise, the receptionist picked up the phone, spoke into it for a few seconds and then informed us that Evgeny Zubarev, the editor in chief of FAN, would be right out to meet us.
Zubarev, who looked to be in his 50s, had close-cropped salt-and-pepper hair and a weary face. He greeted me with a handshake and invited me into his office. We made our way through the turnstiles and signed in with the guards, then took a brief walk down a long hallway to FAN’s two-room office on the first floor. It was unusually quiet for an online news operation that, according to Zubarev, had a staff of 40 people. The newsroom was equipped for a sizable team, with about a dozen identical black desktop computers sitting on identical brown laminate desks, but only two young reporters sat at them. The shades were drawn and the furniture looked just barely unpacked.
As we sat at Zubarev’s desk, I told him about the articles I’d read accusing FAN of being a Kremlin propaganda outfit. He shook his head in indignation. He turned to his computer and brought up FAN’s website, pointing to the masthead and the certificate number that showed FAN was an officially registered Russian mass-media organization. “FAN is a news agency,” he declared. It had stringers and reporters in Ukraine, and in many former Soviet states; they did original reporting, sometimes at great personal risk. Zubarev himself was a veteran journalist who covered the annexation of Crimea for the Russian news agency Rosbalt before joining FAN. But ever since reports linked him to the Internet Research Agency, he had faced questions about his integrity.
“We understand being in this building may discredit us, but we can’t afford to move at the moment,” Zubarev said with a sigh. “So we have to face the situation where reporters like you, Mr. Chen, come in here and ask us questions every day.”
Zubarev said he believed that he and FAN were victims of a smear campaign. I asked him who would do such a thing.
“Listen, that’s my position, not a confirmed fact,” he said. “It’s possible that there are some business interests, I don’t know. Maybe it’s an attack on our investors.” But when I asked who those investors were, he declined to comment. “I can’t discuss the identities of investors,” he said. “That’s in my contract.”
I left St. Petersburg on April 28. One day later, FAN published an article with the headline “What Does a New York Times Journalist Have in Common With a Nazi From St. Petersburg?” The story detailed a mysterious meeting in St. Petersburg between a New York Times journalist — me — and a neo-Nazi. Its lead image was a photo of a skinhead giving an enthusiastic Nazi salute. But it was not just any skinhead. It was the skinhead whom Katarina Aistova brought to our meeting and introduced to me as her brother. As I learned from reading the article, Aistova’s “brother” was in fact a notorious neo-Nazi named Alexei Maximov.
The article explained that Maximov, who goes by the nickname Fly, is a member of Totenkopf, a prominent skinhead group in St. Petersburg. He reportedly served nine years in prison for stabbing a man to death. Just a month before I met him, Maximov again made headlines when, during an investigation into beatings of immigrants around St. Petersburg, the police found weaponry and Nazi paraphernalia in his apartment.
The story made no mention of Katarina Aistova or the Internet Research Agency. Instead, the article claimed I met with Maximov because I wanted his help in creating a provocation against Russia. Maximov told FAN that I requested to meet him because I was “very keenly interested in sentiment among Russian nationalists.” He continued: “He evidently needed stories about how the murderous Kremlin regime persecutes free Russian people. It’s not the first time I’ve come across such requests on the part of Western journalists, but I’m not going to help them with this. Many want to see in Russian nationalists a ‘fifth column,’ which will function on orders from the West and sweep away the Kremlin.” Apparently I was trying to foment a mini-Euromaidan, right there in St. Petersburg.
The article was illustrated with photos of my meeting with Aistova and Maximov. One photo appears to have been shot surreptitiously through the restaurant window while we sat and talked. The point of view is such that Aistova is barely visible; indeed, at first glance, I seem to be having a friendly chat with a skinhead over a cup of coffee. Another photo, this one taken outside the restaurant, somehow makes me look deep in conversation with Maximov, even though I distinctly recall that Aistova was standing between us.
I had to admire the brazenness of the scheme. I remembered how, at the restaurant, Aistova had sat next to me so I had to twist around to talk to her, while Maximov sat silently across from us. Apparently they had arranged themselves so it could appear, from the right perspective, that I was meeting Maximov alone. I emailed Aistova to ask her to explain what happened. She responded only: “I would also like you to explain yourself and the situation!!” (A few weeks later, when I tried calling her by phone, she pretended I had the wrong number.)
Over the course of a few days, the sensational story circulated among a network of small pro-Kremlin blogs. In fact, the FAN story itself had been aggregated from another pro-Kremlin news site called People’s News, which Andrei Soshnikov, the Moi Raion journalist, has reported also operates out of 55 Savushkina. As it spread, it mutated to become even more alarming. One website suggested I was working for the C.I.A.; another, the National Security Agency. A YouTube channel called Russia Today — not the well-known state television channel but a knockoff — posted a slick video about the meeting, set to a pounding dubstep soundtrack. Disconcertingly, it included a photo of me leaving my hotel. The video currently has more than 60,000 views. Many of those views were a result of a familiar pattern of social-media promotion: Dozens of trolls on Twitter began tweeting links to the video using the hashtag #ВербовкаНацистов — “Recruitment of Nazis.” The hashtag trended on Russian Twitter.
After recovering from the initial shock, I began to track the campaign against me. I had practice, after all, from my months spent on the trail of the Internet Research Agency. I Googled the various Russian spellings of my name every hour to catch the latest posts as soon as they surfaced on LiveJournal and VKontakte. I searched Twitter for the URL of the YouTube video to catch every post.
A few days later, Soshnikov chatted with me on Skype. “Did you see an article about you on FAN?” he asked. “They know you are going to publish a loud article, so they are trying to make you look stupid in front of the Russian audience.”
I explained the setup, and as I did I began to feel a nagging paranoia. The more I explained, the more absurd my own words seemed — the more they seemed like exactly the sort of elaborate alibi a C.I.A. agent might concoct once his cover was blown. The trolls had done the only thing they knew how to do, but this time they had done it well. They had gotten into my head.
Correction: June 21, 2015 An article on June 7 about Russian Internet ‘‘trolls’’ referred incorrectly to the Internet platform Yandex. It was subjected to political pressure, but it was not brought under the control of Kremlin allies.
Adrian Chen is a New York-based writer whose work has appeared in Wired, New York magazine, and The New York Times. He is a contributing editor for The New Inquiry and a founder of I.R.L. Club, a regular gathering for people from the Internet to meet “in real life.” 
