Why Scapegoating Social Media is Uniquely Stupid

Wokeness, Decline, and the State Religion

From “Why the Past 10 Years of American Life Have Been Uniquely Stupid”, by Jonathan Haidt

Haidt begins with two tales: the biblical tale of babel and the liberal story of modernity. This is unknowingly poetic. He tells two myths, believing one of them to be fact. We will leave this irony to the end to judge.

Haidt then presents his enemy center stage: social media. In his telling, Zuckerberg and his team open a Pandora's box by introducing metrics.

Shortly after its “Like” button began to produce data about what best “engaged” its users, Facebook developed algorithms to bring each user the content most likely to generate a “like” or some other interaction, eventually including the “share” as well. Later research showed that posts that trigger emotions––especially anger at out-groups––are the most likely to be shared.

It's remarkable that the like button changes humans to prefer anger. This is remarkable because it's false. Anger is salient from speeches, to posters, to classic television ads. The well-trained eye will catch that Haidt didn't explicitly say that the like button is the causal variable, only a correlated one. The way this is framed, most people would not be able to tell the difference.

This new game encouraged dishonesty and mob dynamics: Users were guided not just by their true preferences but by their past experiences of reward and punishment, and their prediction of how others would react to each new action. One of the engineers at Twitter who had worked on the “Retweet” button later revealed that he regretted his contribution because it had made Twitter a nastier place. As he watched Twitter mobs forming through the use of the new tool, he thought to himself, “We might have just handed a 4-year-old a loaded weapon.”

Social scientists like Haidt have a term for this effect: moral panics, coined in 1972. Haidt is clear that this is a historical problem.

It was just this kind of twitchy and explosive spread of anger that James Madison had tried to protect us from as he was drafting the U.S. Constitution. The Framers of the Constitution were excellent social psychologists. They knew that democracy had an Achilles’ heel because it depended on the collective judgment of the people, and democratic communities are subject to “the turbulency and weakness of unruly passions.” The key to designing a sustainable republic, therefore, was to build in mechanisms to slow things down, cool passions, require compromise, and give leaders some insulation from the mania of the moment while still holding them accountable to the people periodically, on Election Day.

Madison makes a crucial insight about human nature. To me, it begins to make the case for a counternarrative about human nature itself. People can often be short-tempered, envious, petty, prideful, or simply wrong. They might impose these vices on others believing them to be virtues. They've done so throughout history. The founders wrote checks and balances into the constitution for this very reason. They understood that some people are better decision-makers than others. We should remember that piece of wisdom.

Haidt argues that loss of public trust is a failure mode. This is not necessarily a false statement, but it obscures a hidden, arbitrary assumption. Here is a key section:

The most recent Edelman Trust Barometer (an international measure of citizens’ trust in government, business, media, and nongovernmental organizations) showed stable and competent autocracies (China and the United Arab Emirates) at the top of the list, while contentious democracies such as the United States, the United Kingdom, Spain, and South Korea scored near the bottom (albeit above Russia).

Recent academic studies suggest that social media is indeed corrosive to trust in governments, news media, and people and institutions in general. A working paper that offers the most comprehensive review of the research, led by the social scientists Philipp Lorenz-Spreen and Lisa Oswald, concludes that “the large majority of reported associations between digital media use and trust appear to be detrimental for democracy.”

I like this section because it's a plausible narrative that I wouldn't fault anyone for believing. Let me give you another plausible narrative that I wouldn't fault anyone for believing:

The most recent Edelman Trust Barometer (an international measure of citizens’ trust in government, business, media, and nongovernmental organizations) showed stable and competent autocracies (China and the United Arab Emirates) at the bottom of the list, while contentious democracies such as the United States, the United Kingdom, Spain, and South Korea scored near the top.

Recent academic studies suggest that social media is indeed supportive to representation in governments, news media, and people and institutions in general. A working paper that offers the most comprehensive review of the research, led by the social scientists Philipp Lorenz-Spreen and Lisa Oswald, concludes that “the large majority of reported associations between digital media use and trust appear to be beneficial for democracy.”

I just gave you a completely valid interpretation of the complete reverse of the data. You might notice that both narratives paint the U.S. in a good light. The latter assumes that the data reveals something true about the state, while the former assumes that it reveals something false. The point here is that a loss in trust is agnostic to whether it is or is not deserved. This entire section relies on the idea that the distrust caused by social media is unjustified. There is no evidence whatsoever to suggest this.

The following section is well done. In the first half, Haidt makes the case that polarization makes accomplishing things more difficult. I don't take objection to the correlation and he gets his facts right. I have more complicated arguments on why institutional decline is inevitable, including in low-polarization and zero-polarization (single party) states which I think matches the world better, but I don't think Haidt says anything false or deceptive here. In the second half, Haidt compellingly compares social media influence to dart guns.

What changed in the 2010s? Let’s revisit that Twitter engineer’s metaphor of handing a loaded gun to a 4-year-old. A mean tweet doesn’t kill anyone; it is an attempt to shame or punish someone publicly while broadcasting one’s own virtue, brilliance, or tribal loyalties. It’s more a dart than a bullet, causing pain but no fatalities. Even so, from 2009 to 2012, Facebook and Twitter passed out roughly 1 billion dart guns globally. We’ve been shooting one another ever since.

…

First, the dart guns of social media give more power to trolls and provocateurs while silencing good citizens.

…

Second, the dart guns of social media give more power and voice to the political extremes while reducing the power and voice of the moderate majority.

…

Finally, by giving everyone a dart gun, social media deputizes everyone to administer justice with no due process.

Any rational participants should be genuinely concerned that institutions take these complaints seriously. The mob is now able to coordinate social attacks at an unprecedented scale. Haidt pinpoints a problem in the status quo.

Haidt coins a term in the next section: structural stupidity. This is an excellent descriptor of the category of problem, but Haidt uses it incorrectly.

In his book The Constitution of Knowledge, Jonathan Rauch describes the historical breakthrough in which Western societies developed an “epistemic operating system”—that is, a set of institutions for generating knowledge from the interactions of biased and cognitively flawed individuals. English law developed the adversarial system so that biased advocates could present both sides of a case to an impartial jury. Newspapers full of lies evolved into professional journalistic enterprises, with norms that required seeking out multiple sides of a story, followed by editorial review, followed by fact-checking. Universities evolved from cloistered medieval institutions into research powerhouses, creating a structure in which scholars put forth evidence-backed claims with the knowledge that other scholars around the world would be motivated to gain prestige by finding contrary evidence.

…

Part of America’s greatness in the 20th century came from having developed the most capable, vibrant, and productive network of knowledge-producing institutions in all of human history, linking together the world’s best universities, private companies that turned scientific advances into life-changing consumer products, and government agencies that supported scientific research and led the collaboration that put people on the moon.

…

[Institutions] got stupider en masse because social media instilled in their members a chronic fear of getting darted. The shift was most pronounced in universities, scholarly associations, creative industries, and political organizations at every level (national, state, and local), and it was so pervasive that it established new behavioral norms backed by new policies seemingly overnight. The new omnipresence of enhanced-virality social media meant that a single word uttered by a professor, leader, or journalist, even if spoken with positive intent, could lead to a social-media firestorm, triggering an immediate dismissal or a drawn-out investigation by the institution. Participants in our key institutions began self-censoring to an unhealthy degree, holding back critiques of policies and ideas—even those presented in class by their students—that they believed to be ill-supported or wrong.

Many of these examples are simply ahistorical. Universities have improved in their scientifical and technological output, but it was following a very specific period of time. "What happened in that time?" you might ask. Well, it was a few years after the start of the cold war. The Soviets had just launched the first satellite into space and it seemed like they were just smarter than us. Backed into a corner, U.S. universities implemented g-loaded standardized tests to recruit the most competent of students to become our future technologists. The academic science system boomed. The decline in academia started in the seventies and accelerated to the present day. The replication crisis, which pushed junk papers with tampered methodologies, peaked a few years before 2010, well before the popularity of social media. This obviously does not track with social media, but does track with affirmative action, "hollistic admissions", and the campaign against standardized tests. Simultaneously, the university expanded, which meant that academia as a whole had to open up to lower iq students even if they had kept the same rigor of assessment. Proliferation of critical theory and radical social justice follows the exact same historical line.

Government agencies abandoned the g-loaded civil service exam in the sixties and seventies and have slumped since. Private companies, on the other hand, continue to be the most productive institutions in American life. This is particularly pronounced in technology and finance, areas known to have the highest iq employees. Companies aren't immune from media pressure, or "darting" as Haidt puts it. If we followed Haidt's argument, we would expect that startups, which have little brand presence and are most vulnerable to media attacks, are most susceptible to darting. This is the opposite of the case. The most notable examples of resistance come from startups like coinbase or shopify, while legacy companies succumb the most. 

Haidt critiques the right for having stupid beliefs:

The stupidity on the right is most visible in the many conspiracy theories spreading across right-wing media and now into Congress. “Pizzagate,” QAnon, the belief that vaccines contain microchips, the conviction that Donald Trump won reelection—it’s hard to imagine any of these ideas or belief systems reaching the levels that they have without Facebook and Twitter.

This is an example of the golden age fallacy, since conspiracy theories have not significantly increased. I frequently make the point that the most widely believed conspiracy theory, held by 90% of Americans, was that Saddam Hussein was hiding weapons of mass destruction in Iraq. Ironically, this conspiracy theory was spread by the institutions which Haidt and Rauch valorize. From my own article, the Generation of Fraud:

Prior to the Iraq war, over ninety percent of Americans believed Saddam Hussein’s regime possessed weapons of mass destruction. This was not based on any substantive evidence and in hindsight was unequivocally false. The population of the third-largest nation on earth did not come to believe such an absurd, baseless idea from social media influencers, Russian agents or unfettered internet conversations. Instead, they were the last link in a network of unthinking trust, which passed falsehoods down from intelligence operatives to politicians to established media to each and every one of us.

Haidt also criticizes social progressives, which he asserts are categorically different than liberals:

But when the newly viralized social-media platforms gave everyone a dart gun, it was younger progressive activists who did the most shooting, and they aimed a disproportionate number of their darts at these older liberal leaders. Confused and fearful, the leaders rarely challenged the activists or their nonliberal narrative in which life at every institution is an eternal battle among identity groups over a zero-sum pie, and the people on top got there by oppressing the people on the bottom. This new narrative is rigidly egalitarian––focused on equality of outcomes, not of rights or opportunities. It is unconcerned with individual rights.

Do we remember the same Democratic party? The Democratic party was implementing affirmative action, an explicitly equal-outcomes policy. It's judges weaponized the disparate impact legal doctrine, which criminalizes every policy with, you guessed it, different outcomes. Wokeness today is a direct descendent of this ideology which has shaped Democratic policy since the sixties. Extend the history further back to discover that they didn't invent equal outcomes either. Equal outcomes Maoism drove my family from China while starving 60 million. And the sentiment of equal outcomes, which we call envy, long predates Mao or any form of communism.

Haidt goes further:

You can see the stupefaction process most clearly when a person on the left merely points to research that questions or contradicts a favored belief among progressive activists. Someone on Twitter will find a way to associate the dissenter with racism, and others will pile on. For example, in the first week of protests after the killing of George Floyd, some of which included violence, the progressive policy analyst David Shor, then employed by Civis Analytics, tweeted a link to a study showing that violent protests back in the 1960s led to electoral setbacks for the Democrats in nearby counties. Shor was clearly trying to be helpful, but in the ensuing outrage he was accused of “anti-Blackness” and was soon dismissed from his job. (Civis Analytics has denied that the tweet led to Shor’s firing.)

Once again we are not without historical analogues. I call this neo-Lysenkoism, after Soviet 'biologist' Trofim Lysenko for a reason. Myself writing, The Rule of Midwits in Tablet:

Lysenko is the standard bearer for pseudoscience driven by confirmation bias. But his false claims were not at all random or founded in ignorance: Each of his lies formed part of a coherent argument for communist assumptions. He denied that genes exist on the basis that believing they do could prove a “barrier to progress.” He dismissed contrary evidence presented by Western scientists as “tools of imperialist oppressors.”

There is no shortage of this phenomenon in contemporary times. Think of the denial of diversity in human intelligence, of physical differences based on biological sex, of standardized testing as race-neutral, and of the empirical data behind the efficacy of corporate diversity training and measures to combat the “gender pay gap.” With the benefit of hindsight, it is widely known how Lysenkoism ended: The implementation of his “research” in reality helped facilitate the starvation of tens of millions of people from Ukraine to China through manmade famines.

Luckily for me, Haidt finishes this paragraph off with a reference to intelligence, which gives me prime real estate to put my counternarrative in full swing:

American politics is getting ever more ridiculous and dysfunctional not because Americans are getting less intelligent. The problem is structural. Thanks to enhanced-virality social media, dissent is punished within many of our institutions, which means that bad ideas get elevated into official policy.

My counternarrative thesis:

American politics is getting ever more ridiculous and dysfunctional because some institutions are failing to tell the difference between intelligent and unintelligent people. The problem is structural. Thanks to blank-state assumptions that humans have equal intelligence, institutions that select for unintelligent people elevate bad ideas into policy.

Isn't this much simpler? The timeline also matches recent history, unlike Haidt's narrative. The next section is about Haidt's predictions for the future. This is a good test of his theory:

in a 2018 interview, Steve Bannon, the former adviser to Donald Trump, said that the way to deal with the media is “to flood the zone with shit.” He was describing the “firehose of falsehood” tactic pioneered by Russian disinformation programs to keep Americans confused, disoriented, and angry. But back then, in 2018, there was an upper limit to the amount of shit available, because all of it had to be created by a person (other than some low-quality stuff produced by bots).

Now, however, artificial intelligence is close to enabling the limitless spread of highly believable disinformation. The AI program GPT-3 is already so good that you can give it a topic and a tone and it will spit out as many essays as you like, typically with perfect grammar and a surprising level of coherence. In a year or two, when the program is upgraded to GPT-4, it will become far more capable. In a 2020 essay titled “The Supply of Disinformation Will Soon Be Infinite,” Renée DiResta, the research manager at the Stanford Internet Observatory, explained that spreading falsehoods—whether through text, images, or deep-fake videos—will quickly become inconceivably easy. (She co-wrote the essay with GPT-3.)

There are two problems here: quantity and quality. On quantity, Haidt asserts that there is a substantive difference between the quantity of information now and an unlimited amount. The quantity of tweets on a politically salient topic now easily exceeds 10^5. This is a clear underestimate of the total amount of information on that topic. No one is reading all 10^5 pieces of information, just like how they won’t read everything that’s automatically generated.

The argument about quality is a more compelling one. The framing is crucial: is generalized skepticism a problem? The general population is terrible at statistics, and that the greatest cause of disinformation happens when they conflate highly improbable events with probably ones that are worth making national policy over. Funnily enough this includes both anti-vax (vaccine side effects) and woke (police shootings), which Haidt pokes at earlier. High quality deepfakes solves an existing problem. In fairness to Haidt, there will be a period of transition where people do actually believe the deepfakes to be real and will act accordingly, probably implementing a slew of deranged policies which will take time to undo. It may almost be as bad as the Floyd riots or anti-vax deaths. Haidt is right that Russian propagandists may use these techniques to spread disinformation, but along the same lines, they currently cause more damage by amplifying existing tail events. 

This exploit is notable because it can be measured outside of a political valence. Loss aversion is the term for the asymmetric fear of loss. Most people wouldn't take three to two odds on a coin flip. They want at least two to one. This occurs even when there are no political priors, or really priors of any kind aside from wanting resources. Is the most pervasive obstacle to good thinking really confirmation bias, as Haidt suggests earlier? This doesn’t appeal to factor into loss aversion. Considering that most Americans cannot solve simple exponential equations, a fundamental aspect of pandemic risks, it would be difficult to imagine them balancing pandemic tradeoffs properly. Jonathan Haidt is a psychologist; he of all people should know the literature on individual differences in intelligence. 

The solutions section offers little for analysis, as one should expect that the wrong diagnosis leads a doctor to remove the wrong organs. Instead, enjoy the liberal creation myth from his introduction in all its glory:

There is a direction to history and it is toward cooperation at larger scales. We see this trend in biological evolution, in the series of “major transitions” through which multicellular organisms first appeared and then developed new symbiotic relationships. We see it in cultural evolution too, as Robert Wright explained in his 1999 book, Nonzero: The Logic of Human Destiny. Wright showed that history involves a series of transitions, driven by rising population density plus new technologies (writing, roads, the printing press) that created new possibilities for mutually beneficial trade and learning. Zero-sum conflicts—such as the wars of religion that arose as the printing press spread heretical ideas across Europe—were better thought of as temporary setbacks, and sometimes even integral to progress. (Those wars of religion, he argued, made possible the transition to modern nation-states with better-informed citizens.) President Bill Clinton praised Nonzero’s optimistic portrayal of a more cooperative future thanks to continued technological advance.

The early internet of the 1990s, with its chat rooms, message boards, and email, exemplified the Nonzero thesis, as did the first wave of social-media platforms, which launched around 2003. Myspace, Friendster, and Facebook made it easy to connect with friends and strangers to talk about common interests, for free, and at a scale never before imaginable. By 2008, Facebook had emerged as the dominant platform, with more than 100 million monthly users, on its way to roughly 3 billion today. In the first decade of the new century, social media was widely believed to be a boon to democracy. What dictator could impose his will on an interconnected citizenry? What regime could build a wall to keep out the internet?

The high point of techno-democratic optimism was arguably 2011, a year that began with the Arab Spring and ended with the global Occupy movement. That is also when Google Translate became available on virtually all smartphones, so you could say that 2011 was the year that humanity rebuilt the Tower of Babel. We were closer than we had ever been to being “one people,” and we had effectively overcome the curse of division by language. For techno-democratic optimists, it seemed to be only the beginning of what humanity could do.

In February 2012, as he prepared to take Facebook public, Mark Zuckerberg reflected on those extraordinary times and set forth his plans. “Today, our society has reached another tipping point,” he wrote in a letter to investors. Facebook hoped “to rewire the way people spread and consume information.” By giving them “the power to share,” it would help them to “once again transform many of our core institutions and industries.”

In the 10 years since then, Zuckerberg did exactly what he said he would do. He did rewire the way we spread and consume information; he did transform our institutions, and he pushed us past the tipping point. It has not worked out as he expected.

This is truly a wonderful example of the limits of self-deception. Haidt asserts what he believes to be true, follows it up with what Zuckerberg thinks, and then topples the tower with reality. Wonderful. He goes on to kick the pieces of Zuck's dream without realizing that he toppled over his own assumptions too! 

Haidt's ideology is liberalism, which is often conflated with democracy. Liberals use democracy to mean liberalism, which is why they believe that social media undermines democracy. This paragraph is particularly beautiful in illustrating this contradiction:

Recent academic studies suggest that social media is indeed corrosive to trust in governments, news media, and people and institutions in general. A working paper that offers the most comprehensive review of the research, led by the social scientists Philipp Lorenz-Spreen and Lisa Oswald, concludes that “the large majority of reported associations between digital media use and trust appear to be detrimental for democracy.” The literature is complex—some studies show benefits, particularly in less developed democracies—but the review found that, on balance, social media amplifies political polarization; foments populism, especially right-wing populism; and is associated with the spread of misinformation.

Compare this to how James Madison uses democracy, which is correct:

It was just this kind of twitchy and explosive spread of anger that James Madison had tried to protect us from as he was drafting the U.S. Constitution. The Framers of the Constitution were excellent social psychologists. They knew that democracy had an Achilles’ heel because it depended on the collective judgment of the people, and democratic communities are subject to “the turbulency and weakness of unruly passions.” The key to designing a sustainable republic, therefore, was to build in mechanisms to slow things down, cool passions, require compromise, and give leaders some insulation from the mania of the moment while still holding them accountable to the people periodically, on Election Day.

Liberals include in their definition of democracy institutions that were built explicitly to restrict democracy. This is because their definition of democracy is not democracy, it is liberalism. The truth is that social media is toxic to liberalism, not because it changes people, but because it exposes the extremes of psychological traits. Believing that human intelligence is uniform is like believing population structure is uniform: a simple mistake that will completely destroy your understanding of reality. Social media shines the light not only on failed liberal institutions, but on the very assumptions which caused them to fail. 

Ultimately, the babel analogy is apt. We have put all our effort into building up the false idol of liberalism. As a consequence, we have split into incompatible political tongues. But the story doesn’t end there. By splitting them from their monomaniacal pursuit, God frees man to discover truths he never would have sought. We need difference and experimentation. We need not unity under a single tower, but rather the freedom to discover anew.