This past October, Jane Mayer reviewed a book by Kathleen Hall Jamieson, a professor of communications at the University of Pennsylvania, titled “Cyberwar: How Russian Hackers and Trolls Helped Elect a President—What We Don’t, Can’t, and Do Know.”
Jamieson’s book is a painstaking analysis of Russia’s online activities in the months leading up to the election of Donald Trump.
Russian operatives, Jamieson reports, created a Facebook page for a nonexistent civil-rights group called Blacktivist, which shared divisive memes and videos and was more popular than the official Black Lives Matter page.
Twitter hosted more than fifty thousand Russian accounts that pushed extremist views.
The Russians hoped to influence individual voters, but they also wanted to reach journalists.
Jonathan Albright, a media scholar at Columbia University, told Mayer that the Russian campaign “amplified wedge issues beyond authentic reach through social media, which then magnified media coverage.”
The plan, essentially, was to use virality to hijack journalism and, thereby, to shift our national conversation in directions that favored Trump.
In Jamieson’s view, Russia’s social-media campaign decisively influenced the election.
(The Russians, she argues, also succeeded in using WikiLeaks to set priorities for the F.B.I.)
This was achieved, in part, through the “agenda-setting effect” that social media has on journalists.
Russian operatives excelled, for example, at gaming Twitter’s list of trending hashtags.
Shortly after the fund-raising event at which Hillary Clinton said that half of Trump’s supporters belonged in a “basket of deplorables,” Russian bots and trolls joined with real American citizens to make #basketofdeplorables a trending hashtag.
Journalists tend to interpret the fact that a topic is trending as a sign of its newsworthiness—“By Saturday morning, #BasketofDeplorables was trending on Twitter,” a Times story noted—and editors assigned articles about Clinton’s gaffe.
Since more than seventy per cent of journalists check Twitter to monitor their competitors, a snowball effect took hold, and stories multiplied.
Russia’s victory was twofold.
The media’s focus on Clinton’s remark increased its inflammatory reach, providing the Trump campaign with an ironic slogan.
Meanwhile, articles about self-identified “deplorables” exacted an opportunity cost, as readers invested time in absorbing them and journalists left other stories unwritten.
I thought of “Cyberwar” this week, while following the saga of the Covington Catholic High School boys.
Late last Friday, a video surfaced that appeared to show a group of them aggressively surrounding and mocking a Native American activist at the Indigenous Peoples March, in Washington, D.C.
(The students were in town for the March for Life, an annual anti-abortion protest that took place on the same day.)
On Twitter, the students were swiftly condemned by many people, including journalists; one student in particular, Nick Sandmann, was singled out for the way he smirked at the activist, Nathan Phillips.
(“Have you ever seen a more punchable face than this kid’s?” the religion scholar and media personality Reza Aslan tweeted.)
Reporters began covering the story, citing the attention the video had received on social media.
Administrators at Covington Catholic apologized and suggested they might expel the students, and the mayor of Covington, Kentucky, condemned their “appalling” behavior.
Then the screw began to turn.
More videos surfaced, revealing that the initial framing of the story had been somewhat misleading: while it was true that the students were maga-hat-wearing boys who had attended an anti-abortion rally, they had not approached Phillips as initially assumed.
Some people who had criticized the students apologized; some of the apologizers received death threats from deranged supporters of the boys.
In a piece called “Don’t Doubt What You Saw With Your Own Eyes,” Laura Wagner, of Deadspin, criticized those who had walked back their criticisms, arguing that they were chasing respectability by distancing themselves from “the reactive social-media mob.”
Think pieces blossomed. Sandmann, whose family had hired a public-relations firm, appeared on the “Today” show.
Meta-think-pieces were assigned, including this one. Ross Douthat, at the Times, wrote the ultimate meta-piece—a dialogue with himself in which he took two positions about the video simultaneously.
It was possible, as this process unfolded, to debate the rights and wrongs of what had happened, scrutinizing each new video for decisive and telling moments.
The sad truth, however, was that no one’s behavior was above reproach.
The boys, using intimidating body language, shouted their school cheers as though they were at a pep rally and not on the steps of the Lincoln Memorial; some seemed to mock Phillips in the way they sang and danced along with his drumming; at times, they resembled soccer hooligans more than defenders of the sanctity of life.
Phillips, meanwhile, mischaracterized the nature of the encounter when he first spoke with the news media—among other things, he told the Washington Post that Sandmann had “blocked my way,” when, in fact, he had been the one who approached Sandmann.
All the while, a group of protesters from the Black Hebrew Israelites, another religious group, insulted everyone, telling the Native Americans, “You worship the buffalo . . . A dead buffalo can’t save you!,” and calling the Covington students “incest babies.”
Watching all these videos—the Times produced an excellent compilation—one can’t help but notice the smallness of the event.
Behind those assembled, an empty National Mall yawns.
Only a few dozen people were actually involved. No violence occurred. The climax of the so-called confrontation was an enigmatic staring contest which, after a few minutes, ended of its own accord.
The image of Sandmann and Phillips facing off was evocative.
But the moment itself, supposedly representative of the state of our union, was actually fairly ordinary—just one of the numberless encounters that happen at any protest where different kinds of people, some of them eccentric, gather and mingle.
Commentators tended to weigh in on the content of the Covington video before looking into the circumstances of its virality.
When they did look into how the video had gone viral, they discovered that it had gained a large part of its momentum from a Twitter account called @2020fight, which, last Friday, tweeted it to forty-one thousand followers with an outraged caption (“This maga loser gleefully bothering a Native American protestor at the Indigenous Peoples March”).
According to CNN, @2020fight’s version of the video was viewed two and a half million times.
On Monday, Twitter suspended the account for violating its terms of service, and many speculated that @2020fight was a Russian plant.
In fact, the account appears to have been run by a Bay Area teacher named Talia, who used a photograph of a Brazilian model as her avatar and tweeted, on average, a hundred and thirty times a day.
Is Talia a real person, or was the account a form of social-media marketing? Does it matter?
Regardless of who tweeted it, the video was boosted by Twitter’s algorithms and interaction model, which gamify opinion by rewarding users for sharing content quickly,
in an emotionally charged way; in the end, it had an “agenda-setting effect” on journalists, who quickly commenced the work of emphasizing the divisions that polarize Americans instead of writing about other, more substantive subjects.
The processes on which cyberwar depends unfold automatically, because they are built into the way our media platforms work.
For around a decade, people who think critically about the media have worried about filter bubbles—algorithmic or social structures of information flow that help us see only the news that we want to see.
Filter bubbles make it easy to ignore information that could change our views.
But the Covington story is an example of a different problem. It’s a story that’s disproportionately talked about and hard to avoid.
It’s relatively inconsequential, but also inescapable. There is no bubble strong enough to keep it out.
The Covington saga isn’t fake news, strictly speaking. The events on the Mall really happened; what’s more, the surrounding story raises many questions of broad, genuine interest.
How much should we hold teen-agers accountable for their political views?
Would a group of nonwhite demonstrators have been permitted to behave as the Covington boys did?
What is the moral status of Catholicism, and of socially conservative religious institutions generally? (What if the boys had been students at a Jewish or Muslim school?)
How reactive should journalists be? These subjects are interesting to debate, as are the reputations of Sandmann and Phillips. All of this lends the Covington video a kind of moral momentum.
As more people weigh in, the momentum builds.
It would be wrong, however, to take the moral interest of the Covington video at face value. To the extent that the video raises interesting questions, it does so in a particular way.
A reality-television show such as TLC’s “90 Day Fiancé,” which follows lovelorn Americans as they try to spark romance with green-card-hungry foreigners, also raises many interesting questions about global inequality, immigration policy, and the transactional aspects of our intimate relationships.
But it does so within the frame of reality TV—which has its own mercantile, aesthetic, and ideological biases—and it would be unwise to take the insights gained from it and apply them, uncritically, to the rest of life.
Viral video is similarly enframed.
Just as a reality show will receive production funding only if it ticks certain boxes (Is it sexy? Suspenseful? Hobbesian?), so a video will only go viral if it meets certain requirements (Is it inflammatory? Ambiguous?
Does it contain a clickable still image? Are its characters “punchable”?).
Those qualities alone, moreover, don’t guarantee virality: often, memes go viral with the help of bots and algorithms.
Virality is only sometimes an “organic” phenomenon.
For these reasons, it’s equally unwise to use viral video as an occasion for debating serious questions: the viral frame has a distorting effect.
(One might legitimately ask whether all-male schools are good or bad for their students—but the Covington video would be a bad place from which to start that inquiry.)
The conundrum, though, is that, once serious questions are raised, it’s hard—and perhaps even wrong—not to debate them.
It’s in this sense that episodes like Covington are a trap.
The media theorist who best described this problem is Daniel Boorstin, whose book “The Image: A Guide to Pseudo-Events in America,” from 1962, traces the inner logic of artificially generated newsworthiness.
A “pseudo-event,” in Boorstin’s telling, is any happening that exists primarily so that it can be reported upon and debated.
A lavish party held to celebrate the fiftieth birthday of a fashion label is a pseudo-event, because the party exists only so that photographs of it can be circulated.
Boorstin’s more surprising claim is that many serious and genuinely interesting news stories are also based around pseudo-events.
Suppose, he writes, that a reporter asks a government official about a sensitive subject and receives an answer, then asks another official the same question and receives a different one.
A story can now be written about the rift between the officials.
The rift exists only because the reporter asked the questions that he did.
Still, now that it’s been articulated, the officials’ difference of opinion is genuinely newsworthy: a topic of discussion has been created on-demand.
By this method, Boorstin writes, a news outlet can create a “uniform news stream” of “new-fangled content,” all worthy of readers’ time.
Similarly, a politician can stay in the news by staging pseudo-events—leaks, press conferences, and the like—which are both newsworthy and made to order.
Newsworthiness, it turns out, doesn’t have to flow from the intrinsic qualities of events themselves.
It can also be created by someone who knows how to “embroider and dramatize experience in an interesting way.”
Social-media platforms, of course, are specifically designed to encourage such embroidery and dramatization.
The danger of this system, Boorstin observes, is that “pseudo-events spawn other pseudo-events in geometric progression.”
When a Republican congressman faults Barack Obama for not using the term “radical Islamic terrorism,” that’s a pseudo-event.
Obama now has no choice but to explain himself—and when he defends his refusal to use that phrase, he succeeds only in adding another link to the pseudo-event chain.
Pseudo-events multiply, spreading over the media landscape and outnumbering real events, many of which occur locally, and are of less dramatic interest.
Boorstin predicts a future in which pseudo-events make up the preponderance of what we call political life.
“The life in America which I have described,” he concludes, “is a spectator sport in which we ourselves make the props and are the sole performers.”
One can’t ignore the performance, because what people say matters. At the same time, society pays a cost: its attention is artificially directed in some directions rather than others.
Its image of itself is made to shift.
There was a time, in the early days of social media, when mainstream news outlets reported on what happened online as though it weren’t quite real.
Newscasts and newspaper articles would describe debates among bloggers and social-media users from a distance, as discursive curiosities.
Concepts such as “hashtag activism” suggested a divide between online culture and “real” culture.
Critics of this trope pointed out that the people engaged in online debates weren’t basement-dwelling shut-ins but citizens expressing their opinions.
Many belonged to groups that were underrepresented in the media; barred from articulating their views in the opinion pages, they did so on the Web.
In fact, the critics argued, the online world was the real world.
At the very least, it was more real, or more unfiltered, than the walled gardens maintained by the editors and writers at newspapers and magazines.
Today, the picture looks quite different.
Late last year, in New York magazine, Max Read, a former editor-in-chief of Gawker, published an article called “How Much of the Internet Is Fake?
Turns Out, a Lot of It, Actually.”
It’s a catalogue of the fabrications—fake readers, reviewers, personas, messages, images, videos, companies, social-media users, and so on—that have come to characterize online life.
Some of the Internet’s fakeness stems from malign manipulation, as when a manufacturer pays for fake reviews of its products on Amazon.
But much of it flows from the simple fact that Web sites must sort and classify the vast amounts of content they contain.
It’s possible that the total corpus of all the tweets in the world is a reasonable facsimile of what’s on our collective mind.
But we see only a curated collection of those tweets; the ones that float to the top have been through a sorting mechanism, and that sorting mechanism favors virality.
In theory, we could ignore the viral stories.
But, as soon as they are noticed and talked about, viral stories become pseudo-events, gaining in newsworthiness as the debate around them thickens.
A real conversation ensues, but it is often Seinfeldian: captivating, but based on nothing.
When the dramaturgical or rhetorical interest of a debate exceeds the interest of the real events that inspired it, that debate becomes a fantasy—an occasion for dramatizing our values, rather than testing them against the real world.
This, in turn, makes our values feel hollow.
Could the “agenda-setting effect” of social media be resisted?
Some platforms, such as Reddit and Hacker News, allow users to “upvote” what they find interesting while “downvoting” content that they feel isn’t worthy of other users’ attention.
In theory, an upvote-downvote scheme gives users, rather than an algorithm, the power to promote and demote content; it is antiviral. Twitter works differently.
Because it has no downvote function, the only way to criticize a tweet is, essentially, to retweet it.
As a result, controversial tweets grow in popularity; the most popular tweets, in turn, are promoted by Twitter’s algorithms, so that they appear in users’ feeds more frequently.
Whereas voting-based platforms employ something like a democratic logic, virality-based platforms are winner-take-all.
It’s possible that, if journalists thought more about the structures of the platforms they frequent, they could choose more democratic ones, making themselves less manipulable.
Still, there is no such thing as a perfect social-media platform.
People upvote and downvote in herds, and for silly reasons.
Bots can vote, too.
We may need to change the way we think.
Instead of seeing virality as a genuine signifier of newsworthiness, we need to see it for what it is—a product.
Covington is the kind of product our social-media platforms sell to us.
Perhaps we should be warier consumers.