In 2017, after the shock of Brexit and then Donald Trump’s election, Christopher Bail, a professor of sociology and public policy at Duke University, set out to study what would happen if you forced people out of their social media echo chambers.
Bail is the director of The Polarization Lab, a team of social scientists, computer scientists and statisticians who study how technology amplifies political divisions. He and his colleagues came up with a simple experiment. As Bail writes in his recent book, “Breaking the Social Media Prism,” they recruited 1,220 Twitter users who identified as either Democrats or Republicans, offering to pay them $11 to follow a particular Twitter account for a month. Although the participants didn’t know it, the Democrats were assigned to follow a bot account that retweeted messages from prominent Republican politicians and thinkers. The Republicans, in turn, followed a bot account that retweeted Democrats.
At the time, a lot of concern about the internet’s role in political polarization revolved around what digital activist Eli Pariser once called filter bubbles, a term for the way an increasingly personalized internet traps people in self-reinforcing information silos. “The echo chamber idea was reaching its kind of apex in terms of its public influence,” Bail told me. “It nicely explained how Trump had won, how Brexit had happened.” Bail’s team wanted to see if getting people to engage with ideas they wouldn’t otherwise encounter might moderate their views.
The opposite happened. “Nobody became more moderate,” Bail said. “Republicans in particular became much more conservative when they followed the Democratic bot, and Democrats became a little bit more liberal.”
Social media platforms have long justified themselves with the idea that connecting people would make the world more open and humane. In offline life, after all, meeting lots of different kinds of people tends to broaden the mind, turning caricatures into complicated individuals. It’s understandable that many once believed the same would be true on the internet.
But it turns out there’s nothing intrinsically good about connection, especially online. On the internet, exposure to people unlike us often makes us hate them, and that hatred increasingly structures our politics. The social corrosion caused by Facebook and other platforms isn’t a side effect of bad management and design decisions. It’s baked into social media itself.
There are many reasons Facebook and the social media companies that came after it are implicated in democratic breakdown, communal violence around the world and cold civil war in America. They are engines for spreading disinformation and algorithmic jet fuel for conspiracy theories. They reward people for expressing anger and contempt with the same sort of dopamine hit you get from playing slot machines.
As the recent Facebook leaks reveal, Mark Zuckerberg has made many immoral and despicable decisions. But even if he were a good and selfless person, Facebook would still probably be socially destructive, just as most other big social media platforms are.
It turns out that in a country as large and diverse as ours, a certain amount of benign neglect of other people’s odd folkways is more conducive to social peace than a constant, in-your-face awareness of clashing sensibilities. Little is gained when people in my corner of Brooklyn gawk at viral images of Christmas cards featuring families armed to the teeth. And people in conservative communities don’t need to hear about it every time San Francisco considers renaming a public school.
Right-wing politics has come to revolve around infuriating imagined liberal observers. It’s as if angry conservatives live with hectoring progressives in their heads all the time. Social media may not have created this mentality, but it badly exacerbates it. After all, there’s no point owning the libs if none are watching.
The value of psychic distance can apply within communities as well as between them. In 2017, Deb Roy, director of the Massachusetts Institute of Technology Center for Constructive Communication and former chief media scientist at Twitter, held informal meetings in small towns to talk to people about social media. Several times, people told him they had given up speaking to neighbors or others in town after seeing them express their opinions online. It was the first time, Roy told me, that he heard directly from people for whom social media “is blocking conversations that otherwise would have been happening just organically.”
Roy believes that the potential for a healthy social media exists — he points to Front Porch Forum, the heavily moderated, highly localized platform for people who live in Vermont. But it’s notable that his best example is something so small, quirky and relatively low-tech. Sure, there are ways of communicating over the internet that don’t promote animosity, but probably not with the platforms that are now dominant. In a country descending into a perpetual state of screeching acrimony, we might be able to tolerate each other more if we heard from each other less.
Michelle Goldberg is a columnist for The New York Times.