EA considerations regarding increasing political polarization

American politics has become increasingly polarized in recent years. During the ongoing George Floyd protests, observers have pointed out that polarization has hit highs not yet seen in modern American history. Whether this trend of increasing polarization will continue is unclear. However, it is at least plausible that the trend is far from over, and therefore, broad picture implications are worth closer attention.

In this post, I will explore my preliminary predictions under a scenario where polarization continues to increase. While I think that full-scale war—on the level of the civil war—is unlikely to happen in the United States, for reasons I will go into below, I find that the most relevant comparison might be the Chinese cultural revolution. Although the comparison may seem exaggerated, it is still important to explore key similarities to the Chinese cultural revolution, and what is happening in the United States.

If the United States were to experience a cultural revolution-like event, it would likely affect nearly all areas of impact that effective altruists care about, and would have profound effects on our ability to produce free open-ended research on controversial issues. Given that many of the ideas that effective altruists discuss—such as genetic enhancement, factory farming abolition, and wild animal suffering—are controversial, it is important to understand how our movement could be undermined in the aftermath of such an event. Furthermore, conformity pressures of the type exhibited in the Chinese cultural revolution could push important threads of research, such as AI alignment research, into undesirable directions.

When discussing topics as explosive as the one in this post, it is important to stay grounded in solid reasoning and evidence, and to avoid the tendency of waging the war rather than understanding the war. Understanding Julia Galef’s scout versus soldier mindset is helpful here. While in this post I am forced to engage in speculation, it is my hope that readers will judge my argument based on its merits alone, rather than assuming that I’m trying to single out or attack a particular “side” of the current political debate.


Political polarization, as measured by political scientists, has clearly gone up in the last 20 years. It is unclear whether this recent trend is unprecedented, however. For example, some political scientists believe that current levels are higher than at any point after the civil war. Others are more skeptical.

For my purposes, it is not too important for my thesis that current rates are unprecedented. As an assumption for this post, I will only analyze scenarios where the rates of polarization continue to rise, until they reach extreme levels. I believe there are currently no good reasons to think that there’s less than, say, a 10% chance that polarization will get much worse. Given even a 10% chance, the effects of extreme polarization deserve scrutiny and analysis.

While effective altruists could just wait to find out whether polarization will get worse, I believe it is important to conduct this research early for two reasons. The first is that it may be possible to install norms in our communities that effectively guard against the most negative effects of polarization, and therefore, the earlier we detect these trends, the more likely we are to install such norms. Secondly, the very nature of increased polarization makes it more likely that future analysis will be affected by political pressures, and therefore early research will be more level-headed.

Academics have already explored the implications of increased polarization for eroding democratic norms (see here, and here). A central result of such research is precisely what one would expect. From Wikipedia,

Pernicious polarization makes compromise, consensus, interaction, and tolerance increasingly costly and tenuous for individuals and political actors on both sides of the divide. Pernicious polarization routinely weakens respect for democratic norms, corrodes basic legislative processes, undermines the nonpartisan nature of the judiciary and fuels public disaffection with political parties. It exacerbates intolerance and discrimination, diminishes societal trust, and increases violence throughout the society. [...] During this process, facts and moral truths increasingly lose their weight, as more people conform to the messages of their own bloc. Social and political actors such as journalists, academics, and politicians either become engaged in partisan storytelling or else incur growing social, political, and economic costs. Electorates lose confidence in public institutions. Support for norms and democracy decline. It becomes increasingly difficult for people to act in a morally principled fashion by appealing to the truth or acting in line with one’s values when it conflicts with one’s party interests.

It is easy to see how a society entrenched in political polarization could make coordination more difficult in a community like ours. As effective altruists, our explicit mission is to find the most impactful cause areas. In doing so, we engage in critical discourse and frequently debate controversial issues. However, if political pressures in the United States become strong enough, it may no longer be possible to speak openly about the issues we think are most important, since people will feel they must increasingly care more about appealing to a party line, rather than seeking truth.

Some effective altruists have already expressed concern about current trends in polarization. For instance, Seth Baum has written about how we might counter politically motivated misinformation about superintelligence, and Wei Dai has expressed his own worries that epistemic conditions are worsening. More research can be found in the Appendix. In contrast to these previous posts, my purpose is to explore political polarization from a broader perspective, while highlighting the sorts of political pressures that might be placed on our community in the near future.

Recent trends in political polarization

Political polarization has been on the rise since the 1990s and may be at an unprecedented level for the post-Civil War US. There are some political scientists who make the further claim that the last decade or so has witnessed the rise of both right-wing populism and far-left socialism in the West. While these claims are somewhat more controversial, it appears that extreme positions have increasingly been discussed in mainstream news and social media.

Polarization on the left

The popularity of the intersectional social justice movement has absolutely erupted over the course of the past decade. While many of the positions that social justice advocates espouse would be considered uncontroversial among EAs, here I will focus on the way in which the movement has affected political discourse. In particular, “calling out” and “deplatforming” people who are insufficiently committed to the social justice cause, or not careful enough with their language, seems to be relatively common as of the last five years or so.

In the wake of the recent killing of George Floyd, I’ve seen multiple people criticized for simply being silent during the riots rather than taking a position. Silence is seen by many as its own form of defection. Speaking personally, I’ve never seen such explicit examples of political conformity pressures in my life than in the last few weeks. Opposition to illegal acts such as rioting and looting is seen by some as racist whataboutism, as it is wrong to criticize the tactics an oppressed group uses to protest their oppression, and this tactic shifts the blame away from the police onto people of color. (This applies even though it wasn’t just black people rioting). Here is an example of a social democrat and Obama campaign member who was fired from his data analysis job merely for citing research that found riots weren’t effective. This kind of action would have been unthinkable a few years ago.

Since most American educational institutions and mainstream media sources lean toward the political left, the left arguably controls the current dominant narrative. Thus it is important to consider how these institutions have reacted to the rise of the intersectional social justice movement. One recent example is that UC Berkeley has put in place a diversity statement requirement for new hires, which according to academics such as biologist Jerry Coyne, philosopher Brian Leiter, and economist John H. Cochrane, constitutes an ideological purity test.

One professor from UCLA was recently put on suspension for denying a request from black students that they receive special final exam accommodation due to the recent protests. This came after a petition calling for his termination received over 20,000 signatures, and described his response as an “extremely insensitive, dismissive, and woefully racist response to his students’ request for empathy and compassion during a time of civil unrest.” While the petition focuses on the particular language used in his email, I encourage readers to read the email themselves to decide whether something like this would have been considered normal even five years ago.

Another recent event worth noting is the rise of the police abolition movement. It’s pretty clear to neutral observers that abolition is poorly thought out. Websites advocating abolition give almost no explanation for how a society without police would enforce laws, despite law enforcement playing a central role of government going back thousands of years. (One exception is that anarcho-capitalist advocates of police abolition often mention private policing as an alternative.) But the important part is that this policy proposal stands less on its merits and more on the power it derives from shaming those who disagree. People who oppose police abolition are regularly derided as racists on some sections of the internet. For a visceral example, see the response that the mayor of Minneapolis received after he stated he didn’t support the “full abolition” of the Minneapolis police department. This is a powerful social pressure for many.

Recently, I noticed that JK Rowling was attacked by a Twitter mob for a tweet that suggested that biological sex is real. As far as I’m aware, the dichotomy of biological sex and gender identity was well-accepted among social justice advocates about 5 years ago, but now prominent LGBT advocate George Takei has tweeted that it’s ignorant and transphobic to “defend” biological sex as a concept. While it’s true that binary biological sex isn’t a perfectly unambiguous and binary categorization, as the existence of intersex people demonstrates, it arguably is a boundary that it makes sense to draw. This example shows that changes in ideology can occur within a very short period of time, and statements that would have been considered moderate or even progressive a decade ago can get one cancelled today.

Closer to home for the rationality and EA communities, population geneticist and blogger Steve Hsu has recently come under attack by a mob of left-wingers who believe that he is sympathetic to the alt-right. You can read more about this on SSC. This demonstrates how effective altruists are at risk of having to self-censor their thoughts to avoid being cancelled.

Polarization on the right

Because most mainstream news organizations, social media companies, and universities are left-leaning, the right-wing has not been able to enforce its hegemony through deplatforming in the same way that the left has. Instead, right-wingers have recently tended to be the ones who push for stronger norms of freedom of speech. Historically, it seems that whichever side is less culturally powerful is more likely to advocate for freedom of speech (since it benefits them). In the past, the left occupied this underdog position, which can be seen for example in the socialist free speech fights of the early 1900s and the anti-war Free Speech Movement of the 1960s. If the right were to hypothetically gain control over the institutions that shape and censor discourse, it seems likely that the left’s and right’s roles in the free speech debate would swap.

According to Pew Research Center data, while the American left has become far more polarized in its values in the last two decades, the right’s views have remained largely stable. Despite this, in practice it appears that the far-right fringe has more cultural relevance than it did in the 90s. The alt-right movement which peaked around the 2017 Unite the Right rally is the most extreme example, but many would cite Donald Trump and UKIP as more moderate examples of the recent rise of right-wing populism.

While Trump’s policies are in some ways more moderate than the traditional Republican platform, he has repeatedly made provocative statements that sparked controversy and backlash, and probably contributed to the recent rise in polarization. A recent example is his tweet, later censored by Twitter, that included the phrase “when the looting starts, the shooting starts.” It is personally clear to me that far-right communities have felt emboldened by Trump’s election in a way that could not have been matched by a hypothetical election of Mitt Romney, or John McCain. The explanation for this seems to be that Donald Trump was endorsed early in the 2016 primaries by far-right extremists such as Richard Spencer, after he became known for making racially insensitive comments about Mexicans. Despite his politically moderate stances, his rhetoric has been widely described as demagoguery, which has led to enormous amounts of media coverage—both positive and negative. The result of this extremely excessive media coverage has been to entrench political biases in the American population.

One particular risk from the right that’s worth worrying about is a scenario in which Trump loses the 2020 election, but disputes the results and refuses to concede. The evidence for Trump refusing to concede comes from a variety of statements he has made in the past. Throughout 2016, Trump had questioned the legitimacy of the election, suggesting that it would be “rigged” in Hillary Clinton’s favor. The day of the election, when asked, Trump did not commit to accepting the results if he lost. Even after winning, he further claimed on Twitter that millions of people had voted illegally for his opponent. Since that time, Trump has repeatedly alleged that Democrats are attempting to rig the election against him. Recently, his allegations have involved claims that mail-in voter fraud will tip the election in favor of Biden, which prompted Twitter to censor/​fact check him for the first time ever.

While a constitutional crisis may appear unlikely, it’s important to note that it is not unprecedented for a presidential candidate to dispute the results of an election. After the 1800 election, losing candidate and outgoing president John Adams refused to be present for the inauguration of Thomas Jefferson. In 1824, Andrew Jackson had alleged that the winner John Quincy Adams had entered a “corrupt bargain” with Henry Clay, the sitting speaker of the House, which historians generally consider to have led to increased political tensions for the next four years. In 1876, the dispute between Rutherford B. Hayes and Samuel Tilden was strong enough that the north ended the era of Reconstruction in order to appease southern voters. Most recently in 2000, Al Gore refused to concede until the Supreme Court ruled against him in Bush v. Gore, roughly a month after election night.

Scott Aaronson recently reviewed legal scholar Lawrence Douglas’s book Will He Go?, which explores this very hypothetical. In his review, Aaronson estimates a 15% chance that a scenario like this will unfold. This Metaculus question puts the probability of Trump contesting the results of the election at 57%, but it’s important to note that he may initially contest the results and change his mind after the official electoral college vote. I would also highlight these other two relevant Metaculus questions. If Trump indeed refuses to concede, it would undoubtedly result in extreme polarization and bring the stability of American democracy into question.

Both sides

Both the political left and the political right have come to view the other side more negatively than in the past. According to Pew Research, 71% of Americans currently view the conflicts between Republicans and Democrats as “very strong” compared to 56% in 2016, and 48% in 2012. In the same survey, Democrats were more likely to see conflicts between rich and poor, black and white people, rural and urban areas, whereas Republicans primarily saw conflict between the parties. Though, both sides agreed that the strongest conflicts were between the parties.

Similar surveys have revealed negative attitudes present between both sides of the political spectrum. For example, one survey showed that 75% of Democrats viewed Republicans as being more closed-minded, and 55% of Republicans viewed Democrats as being more immoral. The same survey showed that the last few years have seen declining ratings in how members of one party view members of the other, with a possible acceleration in 2019.

Historical parallels

Civil War (United States)

The Civil War is an obvious historical parallel, given that it was the United States’ greatest moral and political crisis, resulting from a gradual buildup in tensions between members of two political factions. If something like the Civil War were to happen again in the United States, it would obviously be a big deal that effective altruists should pay attention to. However, I’m not convinced that a civil war-level event is particularly likely.

One difference between the situation in 1860 and today is that the divide over slavery fell very closely along north-south boundaries, while current political polarization does not have sharp geographic boundaries, but is mainly urban-rural. Since urban and rural areas rely critically on each other for resources, it is unlikely that an urban-rural war could be logistically feasible. Furthermore, while some people have expressed separatist aims in the past (such as for Calexit), the United States appears to have a very strong political tradition of suppressing rebellion, and has thus far successfully maintained its status as the world’s oldest major constitutional republic.

According to Adam Przeworski, there is a very strong inverse correlation between national wealth, and political violence. The United States is currently the wealthiest nation in the world, with very high per-capita wealth, and unexceptional (though worsening) income inequality. More broadly, Steven Pinker has extensively documented the long-run decline of violence in the world in his books The Better Angels of Our Nature, and Enlightenment Now, but in case you are skeptical of Pinker’s reliability, Max Roser has also written this page on Our World In Data outlining the same thesis. Although small-scale insurrectionary experiments like the Capitol Hill Autonomous Zone may persist in the coming years, it does not appear particularly likely that large-scale violent rebellion will sweep through the United States any time soon.

Since the Civil War was the ultimate test of the constitutionality of secession, it is very unlikely that legal secession could succeed either.

Cultural Revolution (China)

The Chinese Cultural Revolution (1966–1976) aimed to purge “counter-revolutionary” capitalist ideology from Chinese society and to ensure the dominance of Maoism. Some of the actions taken to achieve this goal included destroying old artifacts/​texts, sending dissenters to jails and re-education camps, and sometimes killing dissenters.

Since some readers may not be familiar with Chinese history, you can consult Appendix 2 for my summary of the main events during the Cultural Revolution.

While this may sound very different from the current situation in the US, there are a number of similarities:

  • The Cultural Revolution targeted traditional values, and received power from young people (mostly students) who cared most about the revolutionary values in China.

  • Actions taken to counter traditional values ranged from renaming streets, to destroying historical sites, books, and artifacts. Similar measures are ongoing in 2020.

  • The leftist thread of thought present in China at the time isn’t that much different from the thread of thought that exists today, though there are obviously clear differences (e.g., in the modern US there is far more focus on marginalized racial groups).

  • The cultural revolution targeted free speech in particular, and that effect has lasted until today in China. There are numerous examples of contemporary leftists in the US downplaying the importance of free speech.

Why effective altruists should care

Let’s suppose that something vaguely resembling a cultural revolution will take place. Obviously this would have a lot of effects on society as a whole, but how would it matter for the effective altruism movement in particular?

One risk is that the EA movement will be criticized for being too white and for not focusing enough on intersectional issues. This could lead to EA leaders frantically trying to signal that they care about diversity and funding more social justice causes regardless of their effectiveness. And if the EA movement fails to take a sufficiently strong stand, it may get labelled as right-wing or counter-revolutionary and lose status among left-wing academia and media outlets. (In fact, similar accusations have already been made, e.g. see this Vox piece which criticizes the movement for not having enough diversity, and this paper linking long-termism to white supremacist ideology (rebuttal). It seems likely that more accusations will be made in the future.)

Another risk is that the field of AI alignment could become politicized. Perhaps AGI safety will become associated with one side of the political aisle and the other side will adopt a stance of skepticism toward the risks of AGI. This is what happened with climate change and to some extent with the COVID-19 pandemic, so it could play out here as well. Seth Baum has written articles about “Superintelligence Skepticism As A Political Tool” and “Countering Superintelligence Misinformation” which cover this issue. Similar to how pandemics were rarely discussed in partisan terms before COVID-19, the current non-partisan discussion of AI alignment seems unlikely to last.

There are several past examples of scientific research being stifled by ideological constraints, with for instance Lysenkoism being responsible for massive famines in the Soviet Union. With AGI, the stakes are potentially astronomical. I’d recommend the piece “Politics is Upstream of AI” for discussion of how political factors could affect AI development, but more research is needed on this topic.

Furthermore, animal farming abolition and wild-animal suffering could become politicized. Zeke Sherman has proposed a scenario in which political correctness impedes research into animal cognition and welfare in “The Future of Animal Consciousness Research”. There are many crucial considerations in the field of animal suffering reduction, such as how much moral weight to assign to beings of different complexity, whether welfare is net-positive or net-negative, and so on. If political dogma prevents us from honestly investigating these questions, the results could be catastrophic for animal welfare.

These are just a handful of examples of how polarization and restrictions on free speech could negatively affect EA discussion, but I hope they are convincing enough to not dismiss this risk as negligible.

What effective altruists can do

Even if we acknowledge that a cultural revolution is plausible, and that it would have implications for EA cause areas, it might still be unclear what exactly can be done to prevent it or to mitigate its harms. Influencing broad political and cultural norms does not appear at first glance to be tractable or neglected. Before dismissing this cause outright, I would suggest that EAs consider the following three points.

The first is that even if we can’t prevent a cultural revolution from occurring, we can reduce its impact on EA members and organizations by encouraging them to relocate outside of the United States or the West. It might be beneficial for effective altruists to be prepared to move to multiple alternative countries as backups should one option undergo a cultural revolution. Even the smaller change of relocating to rural areas may soften the blow of a cultural revolution centralized in cities.

I have not yet determined which developed countries are the best destinations for migration, but right now I believe the strongest candidates may lie in Asia and the Eastern Bloc. Japan may be one good option, since it has high living standards and a favorable diplomatic relationship with the West, while maintaining its own unique cultural and political identity (for example, resisting communism during the twentieth century).

Secondly, effective altruists are disproportionately employed at companies like Google and Facebook. The policies of social media giants can influence discourse norms on the Web and therefore society as a whole. While EAs working at tech giants may not have enough power within the organizational hierarchy to make a meaningful difference, it’s something worth considering. Another way in this vein that EAs could make a difference is by creating or popularizing discussion platforms that promote rational argumentation and mutual understanding instead of divisiveness (related post).

Thirdly, and most importantly, even if I can’t yet come up with a specific bullet-proof intervention, I believe that the importance of the issue necessitates that the effective altruist community take it seriously, e.g. by conducting more research. It would be prudent to get to work as soon as possible for reasons outlined in the background section.


Appendix 1: Previous EA-adjacent writings on this issue

Wei Dai has written about the balance of intelligence vs. virtue signalling and his observation that polarization has resulted in decreased epistemic conditions.

Kyle Bogosian has outlined his perspective on the costs and benefits of speech policing in “Tentative Thoughts on Speech Policing”. He has also written about “Political culture at the edges of Effective Altruism”.

There was an EA Forum debate consisting of two posts: “Making discussions in EA groups inclusive” and “The Importance of Truth-Oriented Discussions in EA”.

Kelly Witwicki has written about how and why effective altruists can make progress in making our movement more diverse.

If there’s more I’m missing, feel free to provide links in the comment section.

Appendix 2: Timeline of the Chinese Cultural Revolution

The Chinese Cultural Revolution (1966–1976) aimed to purge capitalist ideology from Chinese society and to ensure the dominance of Maoism. It followed the 1963 “Socialist Education Movement”, in which intellectuals who opposed Mao were sent to reeducation camps in the countryside.

In May 1966, Mao wrote about the need to identify and purge “counter-revolutionary revisionists” who had supposedly infiltrated the ranks of the Party. Mao recruited students and workers to form the paramilitary Red Guards and other rebel groups. By August, Red Guards in Beijing began to commit a massive slaughter of 1,772 people in what became known as Red August. Over 30,000 homes were ransacked and over 80,000 families were forced to leave the city. This was just the first of many massacres committed at various points during the active phase of the Cultural Revolution.

Around the same time, Vice Chairman Lin Biao announced the goal of eliminating the “Four Olds”: Old Customs, Old Culture, Old Habits, and Old Ideas. Actions taken to achieve this objective ranged from renaming streets, to destroying historical sites (e.g. temples), books, and artifacts. Religious clergy were arrested and sent to camps.

By December, Mao declared an “all-round civil war” to settle the dispute between his radical faction and the more moderate establishment. In 1967, as radical groups seized power from local authorities across many cities in China, a “violent struggle” began between these different factions.

In May 1968, Mao announced his campaign of “Cleansing the Class Ranks”, in which tens of millions of people were sent to reeducation camps and somewhere on the order of a million people died.

The targets of the Revolution consisted of the “Five Black Categories”: landlords, rich farmers, counter-revolutionaries, “bad-influencers”, and rightists. In addition, academics and intellectuals were often persecuted or killed.

In April 1969, Mao declared that the active phase of the Revolution was over, but in actuality it lasted until 1971. The final death toll from the Cultural Revolution is disputed, with upper-range estimates as high as 20 million.