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.

Background

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

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.