The Social Media Upheaval

Glenn Harlan Reynolds

• Publisher: Encounter Books
• Year: 2019
• eBook ISBN: 9781641770842
• Topic: Politics & International Relations
• Subtopic: Political Process
• Index: Politics & International Relations

Social media giants are poisoning our journalism, our politics, our relationships and ultimately our minds.Glenn Reynolds looks at the up and downsides of social media and at proposals for regulation, and offers his own fix that respects free speech while reducing social media's toll.

The information world is undergoing a transformation, and it’s huge.

How huge? Huge enough that its impact is comparable to the introduction of agriculture, with similarly far-reaching effects on human civilization. And like the introduction of agriculture, though its impact may be positive overall, there will be significant costs along with the benefits, and neither the costs nor the benefits will be borne evenly.

In this short book, I will look at the change in communications technology that has taken place over the past couple of centuries, and particularly over the past couple of decades. I will explore both the benefits and the downsides of these changes, and look at what’s likely to come next. I will also look at efforts to ameliorate the downsides through regulation of online speech and other approaches. I will conclude with some suggestions of my own.

The Change

Society seems to be growing steadily crazier. And maybe it doesn’t just seem to be. Maybe it actually is growing crazier. Science-fiction writer Robert Heinlein’s 1930s future history dubbed the early 21^st century “the Crazy Years,” a time when rapid technological and social change would leave people psychologically unmoored and, well, crazy. Today’s society seems to be living up to that prediction. But why?

I recently read James C. Scott’s Against the Grain: A Deep History of the Earliest States, and one of the interesting aspects of the earliest agricultural civilizations is how fragile they were. A bunch of people and their animals would crowd together in a newly formed city, and diseases that weren’t much of a threat when everybody was out hunting and gathering over large areas would suddenly spread like wildfire and depopulate the town almost overnight.

As Scott writes, an early city was more like a (badly run) refugee resettlement camp than a modern urban area, with people thrown together higgledy-piggledy with no real efforts at sanitation or amenities. He observes that “the pioneers who created this historically novel ecology could not possibly have known the disease vectors they were inadvertently unleashing.”

Then I ran across this observation on Twitter: “The Internet is rewiring brains and social relations. Could it be producing a civilizational nervous breakdown?” And I saw another article noting that depression in teens skyrocketed between 2010 and 2015, as smartphones took over. It made me wonder if we’re in the same boat as the Neolithic cities, only for what you might call viruses of the mind: toxic ideas and emotions that spread like wildfire.

Maybe we don’t know the mental-disease vectors that we’re inadvertently unleashing, just as those early civilizations didn’t understand the physical-disease vectors they were promoting. Looking around at today’s society, that certainly seems plausible.

Hunters and gatherers were at far less risk for infectious disease because they didn’t encounter very many new people very often. Their exposure was low, and contact among such bands was sporadic enough that diseases couldn’t spread very fast. Their environment and lifestyle were such that both diseases and ideas spread slowly.

It wasn’t until you crowded thousands or tens of thousands of them, along with their animals, into small dense areas with poor sanitation that disease outbreaks took off. Instead of meeting dozens of new people per year, an urban dweller probably encountered hundreds per day. Diseases that would have affected only a few people at a time as they spread slowly across a continent (or just burned out for lack of new carriers) would now, in these congested urban centers, leap from person to person in a flash. It’s no surprise that the earliest cities often depopulated themselves via epidemics.

Likewise, in recent years we’ve gone from an era when ideas spread comparatively slowly to one in which social media, in particular, allow them to spread like wildfire. A few hundred years ago, ideas spread mainly by word of mouth, or by books, which had to travel physically. Later they spread via newspapers. Now they spread at the speed of light, and are shared almost as quickly, at the click of a mouse.

Sometimes that’s good, when they’re good ideas. But most ideas are probably bad. Maybe we don’t know the mental-disease vectors that we’re inadvertently unleashing, just as those early civilizations didn’t understand the physical-disease vectors they were promoting. Looking around at today’s society, that certainly seems plausible.

It took three things to help control the spread of disease in cities: sanitation, acclimation, and better nutrition. In early cities, after all, people had no idea how diseases spread, something we didn’t fully understand until the late 19^th century. But rule-of-thumb sanitation made things a lot better over time. Also, populations eventually adapted: Diseases became endemic, not epidemic, and usually less severe as people developed immunity. And finally, as Scott notes, surviving disease was always a function of nutrition, with better-nourished populations doing much better than malnourished ones.

Right now, it almost seems as if the social media world was designed to spread viruses of the mind. And that’s probably because it was. While in the earlier days of the Internet ideas spread faster than before, today in the walled gardens of social media outlets like Facebook, Instagram, or especially Twitter, ideas spread much, much faster, and with less time for rumination or consideration, than ever before. And that’s by design, as social media companies use algorithms that promote posts based on “engagement” – which typically means users’ emotional reactions – and “share” buttons allow each user to pass them on to hundreds or thousands of friends, who can then do the same. This repeated sharing and resharing can produce a chain reaction reminiscent of a nuclear reactor with the control rods removed.

As Jaron Lanier writes, “Engagement is not meant to serve any particular purpose other than its own enhancement, and yet the result is an unnatural global amplification of the ‘easy’ emotions, which happen to be the negative ones…. Remember, with old-fashioned advertising you could measure whether a product did better after an ad was run, but now companies are measuring whether individuals changed their behaviors, and the feeds for each person are constantly tweaked to get individual behavior to change…. The scheme I am describing amplifies negative emotions more than positive ones, so it’s more efficient at harming society than at improving it.”

In the newspaper age, a full day passed from one publication to the next, and newspapers had to physically travel from the printing plant to the readers. It was possible to share newspaper articles, but only to a relatively small number of friends. Even with the advent of television and radio, there was usually a similar time, days or at least hours, between programs, and sharing was basically impossible. Ideas spread quickly, by historical standards, but not with anything like the speed of today’s social media.

This increase in speed, along with some other characteristics of social media, has led to calls for stricter speech regulation. In the pages that follow, I will outline some arguments for such regulation, suggest some counterarguments, and then offer some suggested responses of my own.

In short, criticisms of social media as speech fall into several categories: it is too fast, it is too incomplete, it is too emotional, and it is too untrustworthy. All of these claims have some merit.

Too Fast

For most of human history, ideas spread slowly. They could travel no faster than the human being sharing them, and the only way they could be transferred was word of mouth. Some ideas still did well: Homer’s work, for example, spread in exactly this fashion. But it was slow at best, and Homer is perhaps the exception that proves the rule, as comparatively little thinking from the preliterate era survives.

The invention of writing, and later of printing, meant that ideas could spread much more widely. Their speed was still limited by human speeds of travel (though such speed was improving a bit) but circulation no longer required word-of-mouth contact. A book, handbill, or newspaper could reach multiple readers, and once movable type arrived post-Gutenberg it was easy to produce them in comparatively large numbers.

Right now, it almost seems as if the social media world was designed to spread viruses of the mind. And that’s probably because it was.

Thomas Paine’s pamphlet Common Sense was the best-selling work of the Colonial era; it sold 100,000 copies in 1776, the year it was published. Given that the population of the United States was 2.5 million at the time, that’s quite impressive. If we assume, plausibly, given that copies were distributed and read aloud at taverns and other public places, that each copy was perused by five people, that would mean that 20 percent of the American population – and a much higher percentage of the adult population, of course – was exposed to Paine’s work in its first year. This is huge, but, of course, it was also an exceptional case.

The spread of information got faster throughout the 19^th century, as railroads increased the speed of travel dramatically, and the telegraph meant that, for the first time, information could travel long distances much, much faster than a human could. Multiple editions of newspapers in big cities meant that an idea published in a morning edition could, at least in principle, be responded to or repeated in an afternoon paper. And with the invention of radio and television, ideas could be broadcast to many people – not simply, as with the telegraph, from point to point – at the speed of light.

This represented a major increase in speed of transmission, but the speed of interaction was still comparatively slow. Television or radio news broadcasts were scheduled things, not really interactive with one another, and, most importantly, they were broadcasts, information spread outward from a single point. They were important new technologies, for good or for evil (Hitler’s mastery of radio was a major part of his success) but they still gave people time to ruminate and think about what they broadcast. And these broadcasts weren’t prone to near-instantaneous chain reactions.

Even the Internet made less of a difference, at first, than it might have. Originally, Internet news was just newspaper content ported over to a website. “Newsgroups” on Usenet were probably the first truly interactive medium, in which any individual could post items and everyone else could respond instantaneously. Usenet quickly gained a reputation for “flame wars” and the promulgation of conspiracy theories, which in retrospect probably should have been a warning.

Even the “blogosphere” of the early 21^st century, in which independently run blog sites posted items on news and responded both to Big Media stories and to each other, was more like traditional media in some respects than like Usenet or social media. To read content on blogs, readers had to go there. To interact, bloggers had to read each other’s sites and decide to post a response, generally with a link back to the post they were replying to. If you didn’t like a blog you could just ignore it. A story that spread like wildfire through the blogosphere still did so over the better part of a day, not over minutes, and it was typically pretty easy to find the original item and get context, something the culture of blogging encouraged. (As James Lileks wrote, “The link changes everything. When someone derides or exalts a piece, the link lets you examine the thing itself, without interference.” And bloggers often encouraged their readers to follow the link and – to use a popular blogospheric phrase – “read the whole thing.”) In addition, a story’s spreading required at least a modicum of actual thought and consideration on the part of bloggers, who were also constrained, to a greater or lesser degree, by considerations of reputation. Some blogs served as trusted nodes on the blogosphere, and many other bloggers would be reluctant to run with a story that the trusted nodes didn’t believe.

In engineering parlance, the early blogosphere was a “loosely coupled” system, one where changes in one part were not immediately or directly transmitted to others. Loosely coupled systems tend to be resilient, and not very subject to systemic failures, because what happens in one part of the system affects other parts only weakly and slowly.

Tightly coupled systems, on the other hand, where changes affecting one node swiftly affect others, are prone to cascading failures. Usenet was one such system, where an entire newsgroup could be ruined by a spreading “flame war.” If a blogger flamed, people could just ignore the blog; when a Usenet user flamed, others got sucked in until the channel was filled with people yelling at each other. (As Nick Denton wrote, the blogosphere “routes around idiots” in a way that Usenet didn’t, because the blogosphere doesn’t depend on the common channel that a Usenet group did.)

Social media – especially Twitter – is more like Usenet than blogs, but in many ways is worse. Like Usenet, it is tightly coupled. The “retweet,” “comment,” and “like” buttons are immediate. A retweet sends a posting, no matter how angry or misinformed, to all the retweeter’s followers, who can then do the same to their followers, and so on, in a runaway chain reaction. Unlike blogs, little to no thought is required, and in practice very few people even follow the link (if there is one) to “read the whole thing.” According to a study by computer scientists at Columbia University and the French National Institute, 59 percent of people who share a link on social media don’t read the underlying story. (I’m honestly surprised the number isn’t higher.)

As Caitlin Dewey reported in The Washington Post:

Worse, the study finds that these sort of blind peer-to-peer shares are really important in determining what news gets circulated and what just fades off the public radar. So your thoughtless retweets, and those of your friends, are actually shaping our shared political and cultural agendas.

“People are more willing to share an article than read it,” study co-author Arnaud Legout said in a statement. “This is typical of modern information consumption. People form an opinion based on a summary, or a summary of summaries, without making the effort to go deeper.”

To verify that depressing piece of conventional Internet wisdom, Legout and his co-authors collected two data sets: the first, on all tweets containing Bit.ly-shortened links to five major news sources during a one-month period last summer; the second, on all of the clicks attached to that set of shortened links, as logged by Bit.ly, during the same period. After cleaning and collating that data, the researchers basically found themselves with a map to how news goes viral on Twitter.

And that map showed, pretty clearly, that “viral” news is widely shared — but not necessarily, you know, read.

Commenting on this study in Forbes, Jayson DeMers writes:

The circulation of headlines in this way leads to an echo chamber effect. Users are more likely to share headlines that adhere to their pre-existing conceptions, rather than challenging them, and as a result, publishers try to post more headlines along those lines. Social groups regurgitate the same types of posts and content over and over again, leading to a kind of information stagnation. This is one of the most powerful negative repercussions of the blind sharing effect.

This trend also makes it easier for journalists and content publishers to manipulate their audiences—whether they intend to or not. In a headline, one small word change can make a big difference, and even if you report all the real facts in the body of your article, the way you shape a headline can completely transform how users interpret your presentation of information. This is a dangerous and powerful tool.

That’s exactly right. Social media makes people less informed but more partisan.

The “block” and “mute” functions on Twitter, and similar tools on Facebook, etc., are intended to protect against Usenet-style flame wars, but to the extent that they work, they also put people in bubbles of similar thinkers, which tends to encourage the spread of misinformation so l...

Table of contents

• Cover
• Table of Contents
• The Change
• Copyright