Network Contagion Research Institute: helping the state fight political infection left and right

In the opening scene of Costa-Gavras’s classic film Z, about the lead-up to the 1967 military coup in Greece, the chief of police (referred to as the General) addresses a gathering of senior government officials on the “ideological disease” he sees threatening their nation. “It is caused by harmful germs and various parasites,” such as socialism, anarchism, beatniks, and pacifist tendencies. “Infection from ideological mildew” must be “fought preventively” by “the spraying of humans with appropriate mixtures”—indoctrination via schooling, military service, and leafleting the peasantry. In addition, the General declares, opponents of the left—who represent “the healthy parts of our society” or “antibodies”—must be used to “combat and eradicate all diseases.” As the film unfolds, we learn that the disease eradication he has in mind consists of physically breaking up leftist gatherings, beating up anti-war protesters, and murdering their leaders.

I’m repeatedly reminded of this scene when reading the work of the Network Contagion Research Institute, whose very name depicts harmful politics as ideological disease. The NCRI aims to “track and expose the epidemic of virtual deception, manipulation, and hate, as it spreads between social media communities and into the real world.” One of the institute’s “Contagion and Ideology Reports” characterizes disinformation and distrust as “a virus that knows no race, that consumes the poor and rich, that infects and kills people of any political persuasion.” Another report warns that “viral ideologies infect mainstream communities” and urges the use of “information vaccines” as protection. Costa-Gavras’s slightly fictionalized police chief would have been right at home with this discourse.

To be sure, the NCRI has given Costa-Gavras’s General a 21st century upgrade: The think tank doesn’t endorse non-state violence, and the “unhealthy” ideas it aims to stamp out emanate from the right as well as the left. But in other ways, the two are strikingly similar. Like the General, the NCRI is a mouthpiece for the state security apparatus and its commitment to defend the established order. Like the General, the NCRI uses the language of epidemiology to strip threatening ideas of both political content and historical context, reduce people who embrace these ideas to passive vessels, and give its own political project a false veneer of scientific objectivity.

NCRI maps the dissemination of slurs and memes with charts similar to this social network analysis.

Anti-hate politics meets big data

The Network Contagion Research Institute was founded in 2018 and is based at Rutgers University under the directorship of Princeton psychologist and neuroscientist Joel Finkelstein. The institute studies how so-called political extremism spreads and develops via social media. The NCRI hosts webinars, offers a college-level training program in “cyber social network threat detection and strategy,” and has published a series of reports on topics such as COVID-19 disinformation, anti-Asian and anti-Jewish conspiracy theories, the Militia and Boogaloo movements, QAnon, and “militant anarcho-socialist networks.”

The NCRI uses a variety of research techniques, but its special sauce is large-scale quantitative analysis of slurs, memes, and code words. With data sets that consist in some cases of tens of millions of social media posts, institute staff and fellows track the frequency with which specific terms appear on various platforms over time. They correlate these patterns with real-world events, measure the spread of hateful ideas from fringe platforms such as 4chan to mainstream ones such as Twitter, and map associations between different frequently used terms to highlight changes in rhetoric and perhaps ideology. For example, the NCRI’s report on COVID disinformation used such data analysis to argue that in early 2021 conspiracist opposition to vaccines and public health restrictions was being subsumed into a larger, overarching conspiracy theory about a tyrannical New World Order government—and also that anti-vaccine protests tended to occur in counties where intimidation was used against Black Lives Matter protesters.

I’m not a data scientist, and I’m not going to comment on the NCRI’s quantitative methodologies. Yet despite the institute’s seeming technical sophistication, its underlying analytic framework is quite crude and weak. The NCRI uses the “hate” framework that has been promoted by the Anti-Defamation League, Southern Poverty Law Center, and others. Kay Whitlock offers an incisive critique:

“In U.S. progressive politics the hate frame has four main assumptions: First, that hate is rooted purely in irrational, personal prejudice and fear and loathing of difference. In fact, it’s also rooted in ideologies and supremacy, in a historical and cultural context. Second, that hate is hate, and the specificities don’t matter. Third, that the politics of hate is about that crazy irrational feeling, which is caused by personal prejudice gone amok. In this view, hate is not about structures, not about power hierarchies, not about institutional practice. Finally, that hate is perpetrated by extremists, misfits, and loners who are violating agreed-upon standards of fairness, and that hate violence is unacceptable and abhorrent to respectable society.

“In fact, what is called ‘hate violence’—violence directed at vulnerable and marginalized groups—is not abhorrent to respectable society. On the contrary, respectable society has provided the models, policies, and practices that marginalize people of color, queers, disabled people, and in many respects, women. The hate frame disappears considerations of structural violence and substitutes in their place the idea that there are these crazed extremists, and that’s who we have to go after.”

Hate frame assumptions are integral to the NCRI approach. NCRI draws a neat division between hateful and non-hateful speech, with no concern for the variety of ideologies underlying such speech or the historical context in which it arises. In NCRI reports, for example, you’ll find lots of references to racist expression, but no discussion of the differences and relationships between genocidal white supremacism, Proud Boys-style “western chauvinism,” and Oath Keepers-style color-blind ideology—and certainly no discussion of how all of these are rooted in a system of racial oppression that has always been central to U.S. society.

As Whitlock argues elsewhere, the hate frame also treats violence against oppressed groups as a problem to be solved with more policing and longer prison terms—without addressing the ways that police and prisons are themselves active perpetrators of systemic violence against oppressed groups on a massive scale. This too, is reflected in the NCRI approach, which is largely geared toward bolstering law enforcement. The institute’s report on the Boogaloo meme, for example, urges law enforcement agencies to “develop large scale and data-driven approaches and central information-sharing capacity” to track and analyze Boogaloo-type threats—in other words, embrace the NCRI methodology as their own.

The NCRI’s use of the hate framework is particularly egregious because the institute applies it to the radical left as well as the far right. The NCRI’s report on “militant anarcho-socialist networks” repeatedly uses language that links and equates leftists with far rightists. For example, the report refers to anti-police slogans such as ACAB (All Cops Are Bastards) and FTP (Fuck the Police) as “hateful codewords and memes” —putting them in the same category as calls to gas the Jews. The report claims that leftists—like far rightists—demonize and dehumanize political opponents, promote “classic authoritarian narratives,” and advocate “violent insurgency.” A table summarizing their findings asserts that “Anarcho-Socialist extremists” have displayed all or nearly all the same characteristics as Jihadis and Boogaloo: expressing “apocalyptic beliefs,” “utopian legends/narratives,” and “martyr narratives”; using online propaganda and private or fringe internet forums; organizing armed militias; and carrying out “lone-wolf terror attacks.” The only one they’re unsure about is whether leftists have carried out “cell-like terror attacks.”

The equation of right-wing and left-wing violence is fundamentally dishonest for two reasons, as Kristian Williams has argued. First, rightists in the U.S. have carried out far more terrorist attacks than leftists, as the eminently non-leftist Center for Strategic and International Studies has documented. Second, in Williams’s words, whatever tactical or ethical disagreements we may have with leftist attacks, “there can be no equivalency between the violence of a slave revolt and the violence of a slave master, between the violence of anti-fascists and that of the Atomwaffen Division.” The NCRI report on anarcho-socialists doesn’t acknowledge any of that, but its authors do maintain a figleaf of deniability with a footnote cautioning that “This analysis does not suggest that violence from anarcho-socialist militants has yet become as widespread as an organized Jihadi group nor does it have the death toll or historical reach that right-leaning extremism has in the U.S. However, anarcho-socialist bloodshed has been historically substantial on other continents and Western countries.”

The same report also promotes the bogus claim, which has been made by both conservatives and some liberals, that the mass-based riots and violent anti-police activism that followed George Floyd’s murder in 2020 were instigated by a few leftist agitators. The report asserts that small groups of activists such as the Portland Youth Liberation Front were able to “mobilize lawlessness and violence” through sophisticated use of online communication to call up a “network-enabled mob” in numerous cities simultaneously. In other words, a think tank that claims to be combating the spread of harmful conspiracy theories is itself replicating a classic conspiracist myth that has been used to demonize leftists for generations.

Toward a centrist anti-hate coalition

Although the NCRI is a relative newcomer to the extremist-monitoring field, its institutional credentials and impressive-sounding methodology have given it a prominent “expert” status for major media organs such as the New York Times and Los Angeles Times. The NCRI describes itself as “a neutral and independent third party whose mission it is to track, expose, and combat misinformation, deception, manipulation, and hate across social media channels,” assuring us further that it has “no political agenda, profit motive, or university reporting obligations.” A more honest description—based on its list of staff and advisors—would be that NCRI represents a convergence of academia (mainly psychologists and artificial intelligence experts), big tech (notably Google’s director of research), and security agencies (with current or former people from the U.S. military, Department of Homeland Security, National Security Agency, New York City Police Department, and private firms).

In addition to Rutgers, the NCRI lists “affiliations” with three entities: the Anti-Defamation League, Open Society Foundations, and Charles Koch Foundation. The ADL is one of the most prominent watchdog groups monitoring the U.S. far right, but it’s no friend of the left. The organization has long misused the charge of antisemitism to attack Palestinians, Palestine solidarity activists, anti-racist activists, and others. In the 1990s, it was revealed that the ADL had spied on a wide range of progressive organizations for decades; as recently as 2017 it publicly urged the FBI to spy on antifa groups, a call it later retracted.

The combination of Open Society and Koch foundations is pivotal to the NCRI brand. Open Society (George Soros’s grant-giving network) figures in countless right-wing conspiracy theories while Koch is one of the most hated capitalist names on the left, so by listing the two together the NCRI declares that it transcends political divisions by bringing together staunch liberals and conservatives. Put slightly differently, the combination of Soros and Koch support evokes an attempt to foster a broad—but anti-Trump—coalition within the ruling class. (Contrary to what some leftists have claimed, the Koch network never supported Trump and rejected his positions on both immigration and trade.)

The NCRI’s approach dovetails with centrist efforts to woo hardline conservatives away from Trumpism, as witness the institute’s recruitment of former Republican Congressmember Denver Riggleman to its advisory team. In Congress Riggleman was a member of the arch-conservative Freedom Caucus, but he lost his 2020 re-election bid after officiating at a same-sex wedding. Last month the New York Times profiled Riggleman as a courageous opponent of conspiracy-mongering under the title “One Republican’s Lonely Fight Against a Flood of Disinformation.”

Complementing its recruitment of Riggleman, the NCRI has recruited former leftist Alexander Reid-Ross as a senior research fellow. He is the lead author on the NCRI’s COVID disinformation report and a contributing author on at least one other of the institute’s studies. Reid-Ross, who teaches geography at Portland State University and used to moderate the Earth First! Newswire, has had significant influence on many liberal and leftist antifascists with his 2017 book Against the Fascist Creep and numerous articles on related topics. Although he has raised important issues, such as collusion between sections of the left and fascists, his past work is a mixed bag; one 2017 review of Against the Fascist Creep rightly faulted Reid-Ross for using guilt by association, name dropping, and just plain bad writing. In any case, by signing on with NCRI he has repudiated the left, yet his background helps burnish the NCRI’s image as an inclusive home for anti-“hate” scholars of every persuasion.

Larger trends

The Network Contagion Research Institute’s rise reflects larger trends. One of these is the drive to apply big data analysis to the study of political propaganda and social media. There’s a growing body of academic articles based on such studies, most of which have been published in the past five years, and there are other outfits besides NCRI supporting comparable work, such as the Atlantic Council’s Digital Forensic Research Lab. In principle this approach could yield valuable insights, but its potential is radically compromised when it is based on an analytic framework that shields established systems of power and oppression from critique. Such political bias seems unlikely to change, given the technical and institutional infrastructure required to support big data analysis.

Another trend, in the wake of Trump’s downfall, is the drive by a resurgent centrist establishment to harness anti-bigotry and anti-fascism to its own ends. As Faramarz Farbod recently outlined, the resulting top-down “liberal/centrist anti-fascist discourse” poses a number of dangers: blaming Trump without explaining the conditions that made him popular, reproducing the myth that the United States is a democracy, ignoring the far right’s roots in U.S. society and the establishment’s own complicity in the rise of violent reactionary forces at home and abroad, and expanding the powers of the national security state. The NCRI is rooted firmly in this discourse.

The NCRI’s efforts to lump together far rightist and radical leftist politics into the same “hate” category embodies an important theme of centrist anti-fascism. We see a similar approach in a recent threat assessment report on “domestic violent extremism” by the U.S. director of national intelligence, which President Biden requested shortly after taking office. The DNI’s report divides “domestic violent extremists” into five categories: “Racially or Ethnically Motivated Violent Extremists,” “Animal Rights/Environmental Violent Extremists,” “Abortion-Related Violent Extremists,” “Anti-Government/Anti-Authority Violent Extremists” and all others. Kristian Williams comments:

“The most striking thing about this classification system...is its perverse refusal to divide between left and right, instead grouping opposing sides together under other categories. Right-wing militias, sovereign citizens and anarchists, for example, are all listed under ‘Anti-Government/Anti-Authority Violent Extremists.’ Racist and anti-racist violence is compressed into ‘Racially or Ethnically Motivated Violent Extremists.’

“‘Abortion-Related Violent Extremists’ includes both those ‘in support of pro-life and pro-choice beliefs’—despite the fact that the FBI cannot point to any pro-choice violence that escalated above the level of online threats, while anti-abortion fanatics have murdered 11 people and attempted to kill 26 more since 1993.”

These categories don’t reflect intellectual sloppiness, but rather a deliberate distortion of reality to demonize leftists and protect the established order. It’s an analytic approach we need to expose and critique, along with the Network Contagion Research Institute’s pseudo-objective ideology and the state repression agenda it serves.