In June 2014, Aleksandra Krylova and Anna Bogacheva arrived in the United States on a clandestine mission. Krylova was a high-ranking official at the Internet Research Agency in St. Petersburg, Russia, an ostensibly private company that was connected with Russian intelligence. Bogacheva, her road buddy, a researcher and data cruncher, was more junior. Their trip had been well plotted: a transcontinental itinerary, SIM cards, burner phones, cameras, visas obtained under the pretense of personal travel, and, just in case, evacuation plans.
The women made stops in California, Colorado, Illinois, Louisiana, Michigan, Nevada, New Mexico, New York, and Texas, according to a federal indictment issued years later. Beyond that, their activities are not well known. Their mission, however, is now public knowledge: to gather evidence of conditions in the United States for a project to destabilize its political system and society, using the rather improbable weapon of millions of social-media posts.
In their long conflict with the United States, officials in Russia have many tools of sabotage available to them. But the major investment in the social-media project seemed to reflect a calculation that, of all the vulnerabilities of modern American society, its internal fracturing—countryside against city, niece against uncle, Black against white—was a particular weakness.
Russia’s Internet Research Agency, or IRA, had been founded in 2013 as an industrial troll farm, where workers were paid to write blog posts, comments on news sites, and social-media messages. Late that summer, a job posting appeared online. “Internet operators wanted!” it read, according to the newspaper Novaya Gazeta. “Task: posting comments at profile sites on the Internet, writing thematic posts, blogs, social networks.” Plus: “PAYMENTS EVERY WEEK AND FREE MEALS!!!”
Hundreds of workers toiled in 12-hour shifts at the IRA offices on 55 Savushkina Street. Each had to manage multiple fake accounts and produce message after message—reportedly three posts a day per account if Facebook was their medium, or 50 on Twitter. Managers issued detailed instructions about content and obsessed over page views, likes, and retweets.
In the years ahead, the agency would write more than 6 million tweets, and its posts would attract 76 million engagements on Facebook and 183 million on Instagram. Some posts were outright disinformation; others sought to whip up anger at the truth. But their common aim was to amplify the worst cultural tendencies of an age of division: writing other people off, assuming they would never change their mind, and viewing those who thought differently as needing to be resisted rather than won over.
When the IRA’s project became public knowledge, a simplistic, if seductive, story line grew up around it. “Yes, Russian Trolls Helped Elect Trump: Social media lies have real-world consequences,” read the headline of a Michelle Goldberg column in The New York Times. Aiding Donald Trump was indeed among the IRA’s objectives, but it wasn’t the mission’s focus. “The story of Russian interference was a really damaging crutch for the imagination,” the Russian American writer Masha Gessen told me not long ago. “It was something that allowed us to think about Trump as somebody from outer space—or at least from Russia—as a kind of alien body, but also an alien body from which we’re somehow miraculously going to be liberated.”
In time, a more sobering analysis emerged. The Russian mission, far from dropping something on America from outer space, had been to fertilize behaviors already flourishing on American soil. “The IRA’s goals are to further widen existing divisions in the American public and decrease our faith and trust in institutions that help maintain a strong democracy,” Darren Linvill and Patrick Warren, scholars at Clemson University who became prominent analysts of Russia’s campaign, have written. “The IRA has used Trump—and many other politicians—as vehicles to further these twin goals, but it is not about Trump himself.” A report by the research firm New Knowledge provided to Senate investigators described similar goals: “to undermine citizens’ trust in government, exploit societal fractures, create distrust in the information environment, blur the lines between reality and fiction, undermine trust among communities, and erode confidence in the democratic process.”
When I began to read the posts myself, I saw even more clearly how the Russians had gone about this work. They had done more than fan the flames of division. They had encouraged the view that the basic activity of democratic life—the changing of minds—had become futile. The troll farm wanted Americans to regard people with different views as immovable, brainwashed, disloyal, repulsive. “The IRA knows that in political warfare disgust is a much more powerful tool than anger,” Linvill and Warren wrote. “Anger drives people to the polls; disgust drives countries apart.”
Americans didn’t need outside help to see one another in these ways. The culture of the write-off, of mutual contempt and dismissal, could be found everywhere you looked. If anything, this attitude was a rare point of commonality across left and right. The ease with which the Russian government exploited these tendencies is frightening, but it also, perhaps, points to a way out: If Americans are so easily manipulated in the direction of enmity and sniping and rage, might they also be more open to persuasion than we tend to assume? If Russian trolls could pull us apart, can we bring ourselves back together?
Crystal Johnson is an actual person, a real-estate agent in Georgia. I spoke with her once on the phone. When I explained that I was looking into how her identity had been stolen and weaponized by Russian intelligence, she hung up and stopped answering my calls.
Johnson tweeted occasionally under the handle @CrystalSellsLA. Her profile photo shows a Black woman in her 30s or 40s with short blond hair. She’s smiling widely, dressed crisply in a black blazer and a white shirt. She looks like someone you would trust to find you a home. She posted a combination of real-estate insights and inspirational quotations. “Resale homes sales R up,” she wrote back in 2012. “As we learned from the recent bubble that burst, a healthy housing market puts many pairs of hands to work.” On another occasion: “Good morning! There is so much we have to be thankful for.” Even Heracleitus made a cameo: “The content of your character is your choice. Day by day, what you choose, what you think and what you do is who you become.”
In February of that year, a Twitter account with the handle @Crystal1Johnson began to tweet—and it tweeted precisely what @CrystalSellsLA was tweeting. On the first day of 2013, the real Crystal Johnson wished the world Happy New Year—as did her clone. That would be nearly the end of its mimicry, though. The account went silent for two years. And then suddenly it became one of the most influential accounts operated by the IRA’s troll farm.
The second week of December 2015 was a tense one. Trump, still a relatively new presidential candidate, had proposed “a total and complete shutdown of Muslims entering the United States until our country’s representatives can figure out what the hell is going on.” Political observers started saying that his campaign was more than a curiosity or a carnival, that it recalled the beginnings of some of the most dangerous movements in history. On December 10, @Crystal1Johnson was back in action. “#BlackLivesMatter,” the account declared. @Crystal1Johnson would tweet 11 more times that day, a major increase relative to the real Crystal’s posts, and in this noticeably different vein. “KKK was terrorizing us decades before #ISIS appeared,” it thundered. “But in America #KKK still is legal!!” Crystal1 also weighed in on a television remake of The Wiz, a remix of The Wizard of Oz with an all-Black cast. “So white people see #racism in an all black cast but not when black people are victims of #policebrutality?”
A new Crystal Johnson had emerged, less interested in real-estate advice than in deep-rooted racial injustices. That first day, @Crystal1Johnson received only a handful of likes and appears to have acquired a single follower. But over the next two years, the account sent another 8,000 tweets and garnered more than 56,000 followers, putting it in the top 1 percent of Twitter users globally. Measured by retweets, Crystal1 was the second-most-powerful Twitter user in the entire sprawling Russian effort, with some 3.8 million repostings.
Linvill and Warren, the Clemson scholars, put me on to Crystal1 as an exemplar of the IRA’s left-leaning trolls. But they also recommended that I look into another of the agency’s top performers, its tenth-most-retweeted account—a right-leaning troll named Jenna Abrams.
Jenna had a different set of preoccupations. Two months into tweeting, with more than 6,000 followers, the account posted: “Everyone has a beard now and I wonder, is that #beard trend connected with #ISIS or just a coincidence?” In just a few words, the tweet married contempt for city-dwelling hipsters to a fear of terrorism. It suggested a shadowy nexus of difference; not only were your fellow citizens unlike you, but they might be in cahoots with jihadists.
On another occasion, the account sought to meld the left’s pro-abortion-rights attitudes with its aversion to war: “Liberals are brave enough to kill unborn children, but not brave enough to kill our enemies #LiberalLogic.” It framed protest as dependency: “#TamirRice’s family to receive $6 million from Cleveland. That’s the new era of welfares for the Black people.” And it took a swipe at “social justice warriors”— “A tip for SJWs: not all things’re about sexism or racism, things can be just things, stop turning everything into an argument for equal rights.”
According to the analysis provided to the Senate, the Russians were trying to amplify “a roster of social issues,” among them Black culture; police brutality and the Black Lives Matter movement; the pro-police/Blue Lives Matter movement; anti-refugee content; arguments in favor of Trump and against Hillary Clinton; arguments in favor of Bernie Sanders and against Clinton; Texan culture; Confederate history; Muslim issues; LGBTQ issues; religious rights; and gun rights.
But what seemed to me even more significant than the subject matter was how the trolls talked about these issues. Over and over, they used these topics to suggest to Americans a certain way of looking at one another: as menacing, alien, and, therefore, unchangeable. Crystal1’s tweets shared news stories that implied, not incorrectly, the endemic nature of white racism. But this real problem was sensationalized as a lurid story of irreconcilable identities. “White people can see aliens, Bigfoot, the Loch Ness monster but can’t see racism, oppression or white privilege,” she wrote. And another time: “Awful! White people used Black Babies as Alligator Bait.”
Jenna also turned political disagreements into conflicts over identity—“New study confirmed: Men who are physically strong are more likely to take a right-wing stance, while weaker men support the welfare state.” Here, the politics of redistribution was turned into a difference in virility. Liberal men were just plain lazy, the tweets suggested: “How do you starve Bernie Sanders’ supporters? Just put their food stamps under their work boots.”
The troll farm’s work seemed designed to make people wonder if their fellow citizens were really even their fellow citizens. “Does #Mississippi Gov. follow ISIS example??,” Crystal1 asked. Meanwhile, Jenna tweeted that President Barack Obama was “risking the lives of Americans to bring his sunnis in,” and that “Osama bin Laden’s letter looks more like a … Bernie Sanders speech.”
Again and again, the IRA posts were sending the same message: These people are not to be trusted. They will never change. They are who they are. And who they are is a threat.
We were being conned into thinking even worse of one another than we already did. And it was working.
As tempting as it may be to view the Russian operatives as instigators, their talent was not inventiveness, but rather the faithfulness of their mimicry. Many of their tweets were thoughtless, full of typos, or copied and pasted straight from elsewhere on the internet. But they saw the great American write-off from a distance, recognized its potential, and exploited it.
The political culture of our polarized time, confrontational and dismissive, has many sources: the inflammatory incentives of social media; the cynical manipulations of billionaire-owned news outlets; the growing voice of once-marginalized groups; the very real material crises that beg for solutions; the frustration with how little a more civil, more hopeful politics has delivered in the past; the sense that other people are too invested in their privileges to change and, therefore, that the purpose of politics is to protect yourself from those people, instead of trying to reach them.
For these and other reasons, Americans have grown alienated from an idea central to democratic theory: that you change things by changing minds—by persuading.
As a result, social movements on the left that need to grow to win devote more energy to keeping people out than pulling people in. Many political campaigns seem to focus more on mobilizing sympathetic voters than on winning over skeptics. Leaders who attempt outreach to the unpersuaded are attacked by their own side as sellouts.
Indeed, one of the ironies of our time is that some of the most dangerous and antidemocratic movements have managed to make their causes appear welcoming and make newcomers feel at home, whereas some of the most righteous, inclusive, and just movements give off a feeling of being inaccessible and standoffish.
But it doesn’t have to be this way. Plenty of evidence proves that persuasion remains possible, and tenacious people on the front lines of democratic life are showing how it’s done. A few years ago, as the pandemic began and a cloud of doom rose over the horizon, I began to follow a group of these optimists: activists, educators, political professionals, and, above all, organizers. What they shared was their dissent from the great write-off.
Alicia Garza, a prominent activist in the Black Lives Matter movement, argues that those who want a “woke” future must make space for the “still-waking.” Loretta J. Ross, a reproductive- and racial-justice activist, says we need a prodemocracy movement that relies less on the callout and more on the call-in. I visited a summer camp for families who had adopted children of another race where, in contrast to the well-publicized explosions over critical race theory, parents were sincerely grappling with how to convince white Americans to adopt new racial attitudes while neither alienating them nor watering down the truth. I got to know a cognitive scientist and a cult deprogrammer who each work on combatting disinformation and manipulation, and who explained how the dominant approach to dealing with the victims of phenomena like QAnon is all wrong; they are thinking up what a public-health approach to the disinformation problem would look like.
And I learned a great deal about how confused and complicated and contradicted and, therefore, malleable millions of voters are.
A year ago in Flagstaff, Arizona, I visited the office of an organizing group called LUCHA, or Living United for Change in Arizona. Inside was the managed chaos of activism—an array of folding chairs, hand sanitizer, packets of sugar, a microwave above a mini-fridge. On the walls were inspirational posters: Leadership is action, not position.
The group was pushing for a pathway to citizenship for undocumented immigrants. Their methods included confronting politicians such as Senator Kyrsten Sinema and knocking on the doors of her constituents. In traditional political canvassing, campaigners might knock on supporters’ doors to make sure they have a plan to vote, and quickly move on. LUCHA does something different, called “deep canvassing.” Organizers spend as long as 30 minutes at each door, and the goal is to get people to talk and talk—about why they feel some kind of way about transgender people or undocumented people or minimum-wage workers—while the organizer listens without judgment and builds trust before trying to persuade.
My guide to the process was a young LUCHA organizer named Cesar Torres. He was born in Mexico, the son of a carpenter, and didn’t know he was undocumented until he was 15 or so, when he wanted to get a job and his parents had to tell him the truth. Today he thinks of his role as helping hostile or indifferent voters see the humanity of people like him, and he has been amazed at how often he succeeds.
He told me about one of his most memorable interactions. A woman said, “No, I don’t know any immigrants.” But when he kept digging, she realized, “Oh, well, yeah, my sister’s husband is undocumented, and he got hurt at work. He had an accident. He’s in the ICU, and they have no health care, they can’t get worker’s comp, and they’re struggling.” Torres was able to explain that her brother-in-law was just the kind of person who would benefit from a pathway to citizenship. What struck Torres was how the woman’s hostility to immigrants lay on the surface but, right below it, was the seedling of another view. “The message that I was able to get across to her was ‘When you think of immigrants, sure, you’re thinking of the border crisis or gangs or whatever the media wants to bring up that week. But it’s not that. It’s people like me.’”
What Torres and other deep canvassers are trained to do is conceive of the person in the doorway in a very different manner from how most of us might: as divided not against you, but against themselves. Torres isn’t trying to implant some foreign idea in the minds of the people he speaks with. Rather, he’s trying to pit some things going on inside them against other things going on inside them, to get them to re-rank these things. Yes, you don’t like immigrants, but you like that immigrant you know. Or you don’t favor a pathway to citizenship, but you know what it means to be overlooked and shut out. For canvassers, these dissonances are grist for the persuasive mill.
“My discovery in doing this work was that most people are 60–40 around most things,” Steve Deline, a longtime organizer for LGBTQ rights and a co-founder of the New Conversation Initiative, told me. “If we ask them to plant their flag on one side or the other, if we approach them that way, they’re going to do so, because that’s what makes us feel like rational, thinking humans—having an answer to a tough question. But if we approach people with the idea that it’s normal to have complicated feelings, even if they have a Trump sign on their front yard, even if their public face expresses one thing—if we approach them with the assumption of There’s something more going on underneath, oftentimes we find out that there is.”
If this theory of the 60–40 voter who needs help sorting things through has a patron philosopher, it is Anat Shenker-Osorio, a messaging consultant who is upending many of the left’s long-standing assumptions about persuasion.
I followed her work over the past two years as she advised major, if not widely publicized, projects of political persuasion: first, a quiet campaign that brought together disparate groups across the left to try to ensure as smooth a transition of power as possible in January 2021; and then regular Zoom strategy sessions for organizers, activists, and staffers working to implement the Biden agenda. In these circles, Shenker-Osorio is something of a friendly insurgent, because her basic view is that Democrats have persuasion all wrong.
The dominant view in the party, as she sees it, is: You have your base, so don’t worry about them; reach out to those moderates in the middle, and if you need to water down your ideas somewhat, so be it—that is the price of big-tent living. Shenker-Osorio argues that this approach all too often ends up pleasing no one, leaving the base disillusioned and the moderates merely meh.
The error of this way, by Shenker-Osorio’s lights, is a misconception of what a “moderate” actually is. People associate “moderate” with the middle of the road, the center, but Shenker-Osorio thinks that’s a mistake. When it comes to big issues and policies, moderates are confused, torn, not sure which pole is their pole. Which is different from saying they prefer the mean between the two poles. One way to think of this is, if I offer you a choice between a pizza and a burger, and you can’t pick—you’re an undecided voter!—it doesn’t follow that you want a pizzaburger. Maybe you want a pizzaburger, the mathematical midpoint between a pizza and a burger. More likely, you will ultimately resolve the dilemma and go with a pizza or a burger. Your “moderate” stance was a temporary state—a situation, not an identity.
A better term for moderates, then, might be “persuadables.” Moderate implies a taste for the tempered version of a thing. Persuadable implies malleability.
The ranks of the persuadable change from issue to issue, year to year. But Shenker-Osorio thinks about it as a rule of 20–60–20. When you ask people to rate their support for various issues (as opposed to parties, about which people are far more tribal), a fifth are committed to your side; a fifth are reliably for the opposition; and most people are “moderate,” which is to say their minds are in play. Persuadable voters, she told me, are “the ‘Good Point’ People because they’re like this: ‘Good point. But, yeah, good point. But, also, good point.’ And so they’re capable of agreeing with things that are radioactively conservative, and they are capable of agreeing with things that are progressive.”
The ‘Good Point’ People believe that, yes, raising the minimum wage is essential for helping families survive, and, yes, raising the minimum wage is going to crush small businesses and fuel inflation. They believe that, yes, immigrants enrich our lives, and, yes, immigrants cost us jobs. In a survey of persuadable Minnesota voters with which Shenker-Osorio was involved, one group was asked whether focusing on and talking about race is necessary for societal progress, and 85 percent said yes. Good point! Then another group was asked if focusing on and talking about race doesn't fix anything and in fact makes things worse, and 69 percent said … yes! Good point! The same survey asked whether Black people face greater obstacles to success than white people do, and 74 percent of persuadables said yes. Many of those respondents then joined the 62 percent who answered yes when asked if Black people and Latinos who can’t get ahead were responsible for their own destiny. Good point! But also … good point!
What responses like these tell Shenker-Osorio is that persuadables are hungry for clues from the world about how to think. And so she works to create messages that don’t simply sell policy ideas but also try to subtly teach voters how to think about an issue. The best political appeals, she says, are structured like this: shared value, problem, solution. If you were pushing to increase the minimum wage, for example, you might begin by framing this as a shared value: No matter what we look like or what’s in our wallets, most of us believe that people who work for a living ought to earn a living. If you were getting into police reform, you might launch with Whether we’re Black or white, most of us want to move through our lives and our communities without fearing for ourselves or our loved ones. It could be as simple as No matter our differences, most of us want similar things.
Reporting on this army of persuaders, I began to look differently at those Russian trolls. It seemed to me that there was a faint sliver of hope in the Russian experiment. If those who seek to unravel our society can figure out what moves citizens in this fragmented and confusing time, so, too, can those who wish it well. If Americans can be manipulated, they can also be persuaded.
This essay is adapted from The Persuaders: At the Front Lines of the Fight for Hearts, Minds, and Democracy.
When you buy a book using a link on this page, we receive a commission. Thank you for supporting The Atlantic.