Why We’re Divided: The Inhumanity of Social Media
Why We’re Divided: The Inhumanity of Social Media
How trolls, mass shaming and compassion fatigue are changing our politics
On Saturday, July 13, former president Donald Trump was very nearly assassinated. Were it not for a split-second turn of the head, the bullet fired from the AR-15 of a 20-year old lying prone on a rooftop 130 yards away would have gone right through his skull, instead of grazing the upper part of his right ear. His assassination attempt, televised live for the entire nation to see, shows how history is often decided by a margin of centimeters.
First, it is worth noting that this is a major moment in American history. As another commentator put it in the Wall Street Journal, plenty of U.S. presidents have survived assassination attempts, but none at such a polarized time, when the fate of the nation seems so precarious to so many.
However, the kid who shot Trump does not seem to be the poster-child of progressive extremism that conservatives are painting him out to be: investigators have failed to find a clear motive; he was a registered Republican who made a one-time donation to a progressive cause at age 17, which makes his political leanings, at best, ambiguous; and his phone’s search history shows that, prior to the shooting, he had looked up images of both Trump and Biden. He was a loner all through high school, with a nonexistent social media footprint and a side-interest in guns. No one in his family or school (he had no friends) could have ever foreseen what he did.
Still, the political tensions have been building for a very long time, and Trump’s attempted assassination — whether politically-motivated or not — has startled the nation into a long-overdue moment of self-reflection. How did we get to the point where an assassination attempt on a former president felt not only logical, but expected? How did we come to see the other side as so extreme that we immediately began to point fingers — and, some would say, with good reason?
The truth is that politics has reached a breaking point in American life.
While most partisans oppose using violence to “get the country back on track”, according to a PBS News poll, one in five Americans think it is a viable option (the number is higher for Republicans, at 28%, than for Democrats, at 12%).
The root cause appears to be a growing belief, among both parties, that the other side is not just wrong, but evil. According to a Pew Research Poll, in 2016, 47% of Republicans and 35% of Democrats thought the other side was “immoral”; by 2022, those numbers had jumped to 72% and 63%, respectively. A study by a team of political scientists, based on a survey of 1,067 partisans, found a “robust” correlation between seeing the other side as evil and wanting to circumvent “democratic norms” keep them out of power. In other words, to the extent we moralize politics is the extent to which we become political extremists.
Both parties suffer from a kind of creeping tendency toward black-and-white thinking about each other, which makes undemocratic methods, like intimidation, more likely. And they have become more likely. The last few years have seen a rise in politically-motivated violence, as well as efforts to circumvent the rules and longstanding traditions of our democracy.
Democrats have tried to…
Expand the Supreme Court from nine to 13 seats and, more recently, impose term limits, thereby decreasing the chance of a political majority (this became popular after Trump nominated conservative, pro-life justices Amy Coney Barrett and Brett Kavanaugh to the Supreme Court).
Grant statehood to D.C. and Puerto Rico, thereby expanding the share of Democratic voters.
Foment partisan aggression against Donald Trump, culminating in a hateful political climate in which his attempted assassination surprised no one.
Foment aggression against Supreme Court Justices who were considering overturning Roe v. Wade. Senator Chuck Schumer, addressing a mob of protesters, revealed the two Justices by name and said, “You have released the whirlwind, and you will pay the price. You will not know what hit you if you go forward with these awful decisions.”
Condemn Capitol rioters while ignoring and — in some cases — affirming political violence from the left, as in the “mostly peaceful” protests of 2020.
Republicans have had their fair share of blunders:
Trump supporters tried to stop the certification of Joe Biden’s election victory on January 6th, in a riot that left one protester dead and several hundred police officers injured.
Republican leaders censured state electors who did not side with Trump’s (false) claims of a stolen election and, in many cases, replaced them with Trump-friendly officials.
Despite the fact that state audits found no evidence of widespread voter fraud during the 2020 election, Trump and his supporters continue to maintain, ominously, that the 2024 election results will be “rigged” unless Trump wins.
In 2022, a man wielding a hammer charged into Nancy Pelosi’s house and seriously wounded her husband, intending to kill her.
(Notice that I am not making judgments about which party is better or worse; I am simply listing instances where partisan interests have come before respect for our democratic traditions.)
The vitriol is even trickling down to local government. According to a survey of 112 local public officials, 87% of those surveyed “observed an increase in attacks in recent years, while 81% reported having experienced harassment, threats, and violence themselves” — including verbal harassment, physical assault, and cyberbullying. An Illinois mayor “had his home vandalized and was likened to a Nazi due to his support of President Trump.” A gay assemblymaker received death threats and had a game camera, used for hunting, “mounted outside on a tree and pointed at his house.”
Social media seems to be the epicenter of this rise in vitriol. 79% of officials said they were threatened and harassed on platforms such as Facebook and Twitter. The numbers were lower but still outrageously high — around 65% — for officials experiencing abuse in public forums and public city meetings. The reason for this discrepancy will become clear later in the article.
The columnist David Brooks rightly says that when people do not think well of each other, the only way to “unify” the country is through force.
If others are evil and out to get us, then persuasion is for suckers. If our beliefs are defined by our identities and not individual reason and personal experience, then different Americans are living in different universes, and there is no point in trying to engage in deliberative democracy. You just have to crush them. You have to grab power and control of the institutions and shove your answers down everybody else’s throats.
But it is his second point that is so incisive. Americans are living in “different universes.”
There are two fundamental reasons for this. The first reason is that the current social media landscape encourages low-nuance, extremist characterizations of political opponents, and relies on moral outrage for its profitability — a model that traditional news outlets, like newspapers, have now adopted for themselves. The second and more important reason is that a massive cultural revolution has been underway since the 1960s, away from traditional religious values and toward a secular worldview where the political has replaced the spiritual, leading to generation of people who believe that politics is the main arena where the battles between good and evil are fought. The next three articles in this series will cover the first cause, while the fourth will cover the second.
Social media and the outrage machine
In 2024, the National Academy of Sciences published a brilliant study that sheds light on how social media companies fuel political polarization.
Laying the groundwork for the rest of their argument, the researchers behind the study define social media as a “supernormal stimuli”, in that it creates conditions that do not exist in the natural world. They use fast food as an illustration: humans have always sought fatty, calorically-dense foods in times of famine or economic deprivation, but the overabundance of such foods in modern times makes it so that our appetites are mismatched to the kinds of foods we eat, leading to heart disease, obesity and other health problems.
Similarly, while humans are indeed social creatures, the social dynamics of Twitter and Facebook are the complete opposite of those that defined human life for 99% of our history. As the social psychologist Jonathan Haidt puts it, we are “embodied creatures.” We unconsciously factor into our responses people’s body language, physical touch and facial expressions. We also have an inbuilt preference for what the researchers call “moral content.” We feel compassion for victims and want transgressors to be brought to justice. In fact, negative information tends to have a stronger impact than positive information on our first impressions, according to one neurological study. However, these natural instincts do not scale well with social media.
Social media exploits our preference for moral content by providing it in a quantity and extremity never before seen in human history.
First, social media increases the quantity of moral content, for two reasons. (1) Apps like Twitter and Facebook flood people’s feeds with more stuff— information and sensory experiences — than human beings historically had access to. Social media connects around 5.3 billion people around the world, with the average Facebook user scrolling through 300 feet of content a day (this is roughly equivalent to the height of the Statue of Liberty, or reading every page of the New York Times three times over). (2) In addition to increasing the amount of time people spend scrolling through feeds, through its algorithms, social media increases the amount of moral content people are exposed to.
Social media makes money by selling user data to advertisers, who then use that data to make ads that are guaranteed to garner attention. The more time users spend on social media, the more accurate the data, and the more valuable it is to advertisers. Social platforms must also compete with one other for a share of users’ attention or “engagement.” They thus have a vested interest in increasing the time their users spend on their sites. And due to the way we’re wired, in the attention economy, “moral content often generates the greatest engagement.” The result is that social media algorithms are hardwired to promote moral outrage.
As evidence, the researchers found that tweets with “moral-emotional language” are significantly more likely to be shared than neutral tweets, regardless if they are posted by “laypeople” or “political elites.” “Similarly,” the researchers noted, “news stories that are framed morally receive more shares than neutral stories online.” Consequently, morally outrageous news stories are more likely to proliferate online than fair, balanced ones. Indeed, the researchers note that people are much more likely to learn about an immoral act online “than from print, radio and TV combined.”
Furthermore, “In an online environment saturated with moral transgressions, this could lead people to perceive transgressions even when there are none present… [In experiments] when a target [stimulus] is presented with overabundance, people tend to remember the target event even when it is not there (50).”
The implication is that online nonevents may come to shape perceptions and actions in the real world.
The “overexposure” described above helps to explain why many Trump voters believe that the election was stolen despite evidence to the contrary: all it takes is for one disgruntled user to post a picture of a misplaced ballot for a thousand people to share it with their friends. That engagement then makes the picture more likely to be promoted by social media algorithms, further inflating the threat of “election interference.” This phenomenon may also explain the Tocqueville Paradox: even as racist attitudes and policies have been in steady decline since the 1960s, perceptions of racial tension have steadily gotten worse. A study in the Manhattan Institute described how social media may be fueling the increased perception of racism in society:
Combined with smartphone citizen journalism, social media mean that knowledge of white-police-on-black-suspect violence is more likely to circulate widely, where it can ignite riots and boost the salience of the race question. Thus, even as the number of such incidents is declining, each event is more likely to be captured alive and to possess a higher media multiplier effect.
The implication is that social media makes it possible for a statistical anomaly to have the same influence on public opinion as a broad-based trend.
Second, social media increases the extremity of moral content, which places further strain on pre-existing political tensions.
Social media’s algorithmic preference for moral content not only floods people’s feeds with triggering events; it attracts political extremists while silencing or scaring off political moderates, such as right-of-center conservatives (like me) and old school liberals. If moral transgressions have the highest chance of getting promoted by algorithms, it is only natural that sites like Twitter and Facebook would attract people from the fringe corners of our Republic, including activists, antiracists, cable news pundits, self-consciously edgy billionaires, Proud Boys, QAnon cultists, and members of Antifa.
Such an assumption is consistent with the data. The researchers note:
Most social media content is produced by a small subset of users who tend to be the most ideologically extreme and the most active online (58, 65). Indeed, those who have the most negative feelings about a group or topic are also the most likely to share negative content online (66). This may lead the online environment to be saturated by the extreme content posted by those who, in turn, hold the most extreme opinions.
“This feature of the online world,” furthermore, “can artificially inflate people’s perception of animosity and outrage — creating false norms (49).” False norms are those exaggerated perceptions of reality that make us think the other party is evil, depraved, and beyond hope. Examples of false norms may include calling Trump a “threat to democracy” or branding Democrats “groomers.”
To make matters worse, character limits on microblogging sites like Twitter leave little room for nuance while encouraging impulsive thought-sharing. For evidence, look no further than Elon Musk, the CEO of Twitter (now called X). Although he is not a rabid partisan, his short, provocative, meme-ridden posts are a good example of what attracts attention online. When politicized, this style of communication is highly flammable.
In the hands of partisans, Twitter becomes a veritable factory of short, low-context, emotionally-charged (and oftentimes bitingly snarky) remarks that energize supporters while enraging political opponents and alienating everyone in between. The ensuing cycle of outrage and overreaction makes it so that users only see the extreme side of a political party, while moderate factions are rendered “invisible.” For real-world examples, witness the tweets of Ben Shapiro and Matt Walsh, the Black Lives Matter movement, or the American Civil Liberties Union (which used to stand for free speech, but has become yet another progressive arm of the Democratic party).
Indeed, the researchers note that, online, “People with the most extreme views generate the most content (57).” As evidence, “97% of political posts from Twitter/X come from just 10% of users, meaning that 90% of the population’s political opinions are being represented by less than 3% of posts online.” This is made worse by the fact that users “engage in both homophily where they choose to connect with individuals who are ideologically similar to them (46, 47) — and acrophily where they choose those who share their ideology but are slightly more extreme than them (68).”
In other words, social media algorithms not only attract people from the political extremes, through a process of creating false norms, it forces people into camps — you’re either for the groomers or against the groomers! — which end up radicalizing their members. Enter into social media as a political moderate or independent, and chances are you will emerge five years later as someone who is much more political, and much less reasonable.
The researchers conclude: “In summary, we theorize that the overabundance and extremity of online content lead people’s… moral dispositions to be perpetually triggered.”
This has implications for politicians, who pay close attention to the language and ideas of their constituencies in order to accurately reflect the mood of the moment and win more votes. Thanks to social media, politicians have a direct line to what their voter base is thinking and can more easily use the language of moral outrage to their advantage. However, by using such language, politicians foster a toxic political environment. As one study in the journal of Science noted: “Newt Gingrich and his followers achieved electoral success with strongly moralized language in the 1980s and 1990s, inspiring political elites on both sides to double down on the rhetoric of moral outrage (e.g., “disgraceful,” “shameful”), further exacerbating political sectarianism.” Sound familiar?
Furthermore, politicians may indulge in “strategic extremism”, a term researchers use to describe the growing tendency among politicians to ignore undecided or centrist voters and court the radical wing of their party to solidify their electoral success with that group. When citizens become radicalized (or the extreme opinions of those citizens dominate the discourse, as they often do on social media), politicians, hungry for votes, quickly follow suit, voicing support for inane policies and utopian fantasies that would have been unthinkable only a decade ago. Examples of strategic extremism abound, from Joe Biden’s open embrace of progressive policies after calling himself a “moderate” and Donald Trump’s embrace of the Big Lie, to the growing belief among Democrats that abortion, healthcare and education are “human rights” that the government is responsible for guaranteeing.
But it gets worse.
Compassion fatigue and mass shaming
The researchers of the Academy of Sciences study discovered two more trends running parallel from the way morality is augmented online: compassion fatigue and mass shaming. Both trends are a function of the unprecedented and unnatural scale of online interactions.
Compassion fatigue describes how, due to the scale of morally outrageous content online, users find it easier to “assign blame” than to empathize with the victim. The internet acts as a public square where millions of bystanders can watch immoral acts play out every day, a historically unprecedented situation that strains the limits of our empathy.
The researchers note that, until relatively recently, humans lived in fairly isolated communities, suggesting that our instinct for empathy was designed to function best in a more socially intimate context, with people whom we have come to know. Research backs this up. The brain areas responsible for empathizing with “psychological suffering” take longer to activate than the pain centers of the brain, according to Nicholas Carr’s book, The Shallows. The fast-paced nature of online life has been observed to stunt this process.
Similarly, per the researchers, the sheer volume of moral transgressions online actually has a “numbing effect on observers.” They write: “when there are many victims rather than just a few, people are motivated to disengage from a conflict and not act (90). As the number of victims in a scenario increases, the likelihood that people will take prosocial action like donating money actually goes down (90).” Similarly,
When people are repeatedly exposed to the same information about a moral transgression, they later report that that transgression seems less unethical than a novel transgression (92). This may lead them to feeling that the transgression was “not that bad” and therefore reduce their compassion for a victim.
During the early months of the Russia-Ukraine war, images and videos of charred, dismembered corpses could be found littering the feeds of anyone with even a passing interest in what was going on. This dynamic may explain a rise in callous behavior online, such as when #hitlerwasright trended on Twitter in response to Israel’s growing war with Hamas. When we become desensitized to atrocities, morals go out the window.
Mass shaming is the second trend that characterizes our new, digital morality. When users are too mentally exhausted by their blood-soaked feeds to empathize with the victims, they will resort to shaming the transgressors.
In pre-digital societies societies, punishments were meted out by trusted community members with an eye toward restoring and educating the transgressor. The need for cooperation made it necessary to balance shame with restoration (or what Christians would call restoring a brother in “a spirit of meekness”). In social media, no such balance is needed, nor is it achieved; there is no need to cooperate with online strangers you will likely never meet in real life.
Thus, cancel culture can be reduced to a simple equation: compassion fatigue + millions of bystanders + low-cost interactions = mass shaming. The third variable in the equation explains the disparity between the number of threats directed at local officials in real life versus online: it costs much less to insult someone on social media than it does to insult them to their face. Anonymity provides freedom from repercussions and, by extension, gives people license to be moral monsters online, especially if they create separate troll accounts.
Such factors, the researchers note, “can undercut the traditional social function of cooperation and incentivize activities like public shaming that are disproportionate to the original transgression.” “Punishment in this context,” they add, “focuses on exact retribution instead of rehabilitation or direction.”
Furthermore,
Publicly shaming transgressors may actually increase their negative feelings and resentment towards punishers, rather than guilt over their transgressive actions (136, 137). This may lead transgressors to focus on the proportionality of their transgression compared to the reaction of the public, rather than on changing their behavior (117). This can lead to the continuation or escalation of conflict. Transgressors might even develop communities around these grievances and seek revenge.
This can be seen in the way conservatives became much more distrustful of liberals after several people were “canceled” for committing micro-aggressions during the summer of 2020 (as we will see, 2020 was a major year for polarization). Every time someone gets canceled, those sympathetic to the canceled party become more hostile toward those siding with the punisher, until acts of violence like the one we saw on Saturday feel inevitable.
The threat of mass shaming not only has implications for people’s livelihoods (many of those who were canceled lost their jobs or suffered a fatal blow to their reputation), they also have implications for the justice system. Jurors may be biased toward a controversial defendant because they spend time on social media, or they may be scared to render a fair verdict if it means going against the wishes of the Twitter mob. When we can’t trust institutions like the justice system, we are not far from the abyss.
 | A guest post by
|