On September 30th, The Brown Daily Herald ran an article about an analysis of an alleged Russian disinformation campaign on Twitter. The analysis was performed by two undergraduates, Ethan Fecht from Brown University and Jack Nassetta from George Washington University, as part of a report issued by the James Martin Center for Nonproliferation Studies in Monterey, California. The authors examined the sudden appearance of seemingly fake pro-Trump Twitter accounts after the alleged chemical weapons attack in Douma, Syria, on April 7th, 2018. The Nassetta-Fecht report, totaling 50 pages, claims these accounts were created by the Russian Federation to discourage US intervention in Syria after the attack. Evidently, this supposed disinformation campaign was not enough to prevent a US-British-French coalition from bombing three Syrian government installations on April 14th.
Nassetta and Fecht call the reader’s attention to what they claim is the grander Russian “counternarrative” campaign to “defame Western institutions”, and for this they have received national acclaim. On September 17th, Nassetta and Fecht published an article in The Washington Post which included a “cheat sheet” which can be used to identify Twitter trolls seeking to deceive Americans. Later, The Brown Daily Herald spoke approvingly of the report, one of the authors being a Brown student. The Herald article quotes William Potter, head of the Center for Nonproliferation Studies, who lauded the two students’ “significant contribution” to our understanding of social media.
There would seem to be no questioning the merit of the Nassetta-Fecht report, which, like other recent attempts to detect propaganda and “fake news” used statistical and supposedly objective methods. Searching through a data set of 850,000 tweets, the authors discovered the alleged Russian disinformation campaign by examining graphs of Twitter account creation, tables of word frequency and a very Trumpian “word cloud” with a large “MAGA” in its center. Accordingly, their report uses some technical terminology: the authors conclude that “synthetic actors” displaying “inorganic activity” on Twitter used “counternarrative messaging” to engage with political leaders on a “linear plane”. This is a far cry from the vague and imperfect methods of traditional journalism, relying instead on objective measures, “big data”, and convincing “visualizations”.
Yet, closer inspection reveals that the methods of Nassetta and Fecht are a thin veneer covering what is a remarkably explicit endorsement of American military interventionism. Indeed, the report’s unsubtle premise is that wholehearted support for the anti-government opposition seeking to oust President Bashar al-Assad at any cost is the only sensible and safe resolution to Syria’s beastly war. Until recently, this was essentially the White House line on Syria, a position treated with rolling eyes by serious journalists and basically anyone outside of highly indoctrinated American foreign policy circles. Though claiming to use new and fashionable methods, the Nassetta-Fecht report is actually a very traditional exercise in American foreign policy discourse: a myopic endorsement of US military domination and a disregard for its victims.
“A competition to see who can commit the worst atrocities”
The complex and bloody Syrian civil war began with democratic protests against the government of Bashar al-Assad in the spring of 2011. The democratic character of the opposition had changed radically by the end of 2014, when, according to Patrick Cockburn, Middle East correspondent for The Independent, “it was clear that the Syrian military opposition had become wholly dominated by ISIS and al-Qaeda clones[i].” The jihadist character of the opposition was secretly acknowledged two years earlier in a report of the Defense Intelligence Agency (DIA) which concluded that “AQI [al-Qaeda in Iraq] supported the Syrian opposition from the beginning, both ideologically and through the media” and “AQI conducted a number of operations in several Syrian cities under the name of Jaish al-Nusra.” The report also recognized the support for the opposition from the West and its allies in the Gulf, sentiments echoed by foreign policy commentators like Anne-Marie Slaughter. It was this opposition which these respected commentators assumed would easily topple Assad (in as little as a few weeks according to former Israeli defense minister Ehud Barak).
Yet Assad’s brutal regime held on, and the war, which Cockburn has called a “competition to see who can commit the worst atrocities[ii]”, has endured for seven years and killed at least 500,000. Throughout, Western media coverage has been largely dependent on opposition sources, Western journalists being unable to enter rebel areas for fear of execution, as in the cases of James Foley and Steven Sotloff in 2014. Opposition media, often in the form of YouTube videos, purports to provide an accurate and immediate picture of the events of the war, though it has been credibly accused of focusing on government crimes at the expense of equally horrendous rebel crimes. As a result, coverage of the war has been entirely “one-sided[…] to a degree probably not seen since the First World War,” according to Cockburn, who has been reporting from the Middle East since 1979.
Political sentiment in the US has been particularly one-sided on the issue of chemical weapons. The UN and the Nobel-Prize-winning Organization for the Prohibition of Chemical Weapons (OPCW) have formally accused both the Syrian government and opposition of using chemical weapons, the former, for example, in Khan Shaykhun in 2017 and the latter in Umm Hawsh in 2016 (the full OPCW-UN Joint Investigative Mechanism (JIM) report for these incidents is available here.). An OPCW investigation concluded that large quantities of sarin had been used in Ghouta in 2013 but attributed the attack to neither side, adding that the culprit must have at least had access to the Syrian government’s stockpile of chemical weapons. A more recent incident in Douma on April 7, 2018 has not yet been attributed to either side, though the OPCW released a preliminary report in June. Yet, regardless of the evidence available at the time, the mere accusations of chemical weapons use by the Syrian government in Khan Shaykhun and Douma were enough to prompt military action by the US and its allies, determined as they were to remove Assad.
In the early stages of the war, the opposition managed to capture East Aleppo, the Old City of Homs and even a regional capital, Raqqa. During this period, the declaration of the Caliphate in June 2014 confirmed a prediction of the aforementioned DIA report, which claimed the weakening or ousting of Assad could result in the establishment of a “… Salafist principality in Eastern Syria (Hasaka and Der Zor), and this is exactly what the supporting powers to the opposition [e.g., the West, Turkey and Gulf monarchies] want, in order to isolate the Syrian regime…”. The report also correctly judged that the strengthening of Sunni radicals in their war with the regime could create “the ideal atmosphere for AQI to return to its old pockets in [Iraq]…”, just as ISI would eventually conquer Mosul and much of Anbar Provience during the first half of 2014.
Since 2015, Assad and his Russian ally have slowly retaken most of Syria, yet another outcome predicted by the prescient 2012 DIA report. Consequently, Western hopes of regime change seem decreasingly likely, as are the chances of more explicit intervention like in the disastrous cases of Iraq and Libya. If the US sought to leapfrog from Syria into Iran, it will now have to consider other methods. Today, the only prospects for peace in Syria would seem to involve negotiations between Russia and the United States based on the current territorial holdings of the government and opposition. Assad has all but won and any peace will have to take this into account.
Distorting, distracting, dismaying and dismissing “average westerners”
The fog of war is particularly thick in Syria, and Nassetta and Fecht seek to clarify the situation using statistical methods. From a data set of 850,000 total tweets, the authors analyzed 3,740 tweets which met two criteria: 1) they were replies to Donald Trump’s Twitter handle, (@realDonaldTrump) and 2) they contained “counternarrative” phrases like “chemical attack fake” or “douma false flag”. They then combed through about 10% of these 3,740 tweets and classified the Twitter users into six groups, ranging from “Typical Users” (real people without nefarious Russian agendas), to trolls (avatars of “undeclared state actors” who “[masquerade] as an average Westerner.” The authors spare us from a definition of “average Westerner”.). Nassetta and Fecht simply eyeballed these categories according to heuristics like the grammaticality of the tweets, the “bizarreness” and “randomness” of the Twitter handles, or the “extreme partisan” nature of the tweets.
According to their analysis, about 16-20% of tweets written within ten days after the April 7 Douma incident were “disseminated by a well-coordinated, narrowly focused state actor, almost certainly the Russian Federation.” Many of these Twitter accounts may have been fake and indeed many of them may have been operated by agents of Russian propaganda, though no serious evidence of this claim is provided. Nevertheless, it will be worthwhile to examine Nassetta’s and Fecht’s analysis of the “tactics” employed by the accounts in question. For, according to the Nassetta-Fecht report, the fake accounts used four strategies, conveniently beginning with the letter D, to mislead the “average westerners” among whom they masqueraded: “Distort”, “Distract”, “Dismay” and “Dismiss”. We treat each of the “D”s in turn.
Nassetta and Fecht define the “distort” tactic as “[s]electively including pieces of evidence to enhance your narrative over a competing one. Alternatively, creating evidence that gives a false representation of reality.” The authors only provide one example of this tactic, taken from alleged “synthetic actor” Charles Williams with Twitter handle @Olakunle73 (Fig. 1.). It is unclear from the report which of Nassetta’s and Fecht’s heuristics classified Williams as synthetic. The user suggests that the US, Britain and France are “crooks” for bombing Syrian installations after the incident in Douma, using the hashtag “#TheHague” to insinuate that perhaps Western officials should be held accountable for this strike.
The authors imply this tweet represents a “distortion” because Assad has not “always been transparent in calling for chemical weapons inspections at the alleged attack sites.” Putting aside the fact that in the past Assad has indeed complied with international weapons inspectors and destroyed large stockpiles of chemical weapons, why is the Williams tweet a distortion? First of all, it is true that international law prohibits military actions like that of April 14th except in the case of imminent civilian harm. Given that the OPCW-UN JIM has still not attributed the Douma attack to the Syrian government six months later, it is unlikely that the US-British-French Strikes were legal. Second, international inspectors did arrive in Syria on the day of the US-led bombing. While it is unclear whether the airstrikes did in fact impede their investigation, it certainly could not have helped, and simply served to demonstrate the US coalition’s disregard for due process.
A complicated conflict like the Syrian civil war is precisely where serious, slow, careful proof of each crime is needed most. Both sides have used chemical weapons and at least one of them, the opposition, has demonstrated its skill at social media manipulation. Yet, the Nassetta-Fecht critique of this tweet suggests the authors do not think the mission of the OPCW is very important, at least compared to the more pressing matter of American airstrikes. This opinion is hard to square with the authors’ earlier claim that synthetic actors seek to “defame Western institutions”. Who exactly is doing the defaming?
According to the authors, the “distract” tactic uses “conspiracy theories about an incident to cast doubt on the prevailing narrative” and also introduces “doubt about the credibility of a source.” Nassetta and Fecht give two examples of this tactic. The first (Fig. 2) comes from a “Charles M” with Twitter handle @CharlesMorgn. Again, this tweet is not in response to @realDonaldTrump, so it is unclear why it is included in the study. Charles M notes uncontroversially that the Bush administration lied in the run-up to the Iraq war and calls for proof of Assad’s culpability in the Douma attack. This is not an altogether unreasonable request; the OPCW and UN exist precisely to provide proof of guilt in such incidents.
Nassetta and Fecht regard such tweets as distractions because the tweets assume “[t]he West has previously lied about their justification for wars, [sic] therefore they are likely lying about these chemical attacks.” Yet, this is not exactly the claim made by Charles M, who does not say that US is lying, but simply that no evidence has been provided. The report’s authors evidently consider the search for proof a “distraction,” presumably from more serious business, like bombing. Naturally, the authors never actually address the claim that the United States and indeed all great powers throughout history have lied to justify wars.
A second example of “distraction” comes from Twitter user raffaele, whose handle @raf89230428 has a “high degree of randomness”, according to the authors. The user claims (Fig. 3) that the US is hypocritical if it does not also attack Israel, who illegally used white phosphorous in Gaza during Operation Cast Lead in 2009[iii]. Nassetta and Fecht consider this Tweet a distraction since it “divert[s] attention to another matter,” presumably the matter of the United States’ intentions, which should of course never be examined. One imagines that the authors would also consider the claim that the US lacks credibility unless it bombs Riyadh for its ongoing crimes in Yemen (which are far deadlier than any chemical weapons used in Syria) to be irrelevant as well.
The authors give four examples of the “dismay” tactic, which they define as “[t]rying to intimidate an opponent by warning of the dire consequences of action.” One of these tweets (Fig. 4, top) apparently tried to intimidate “average Westerners” by warning of the possibility of nuclear conflict with Syria’s Russian ally. The US and Russia partially coordinate their military operations in Syria, practicing “deconfliction” to avoid direct conflict and alerting one another about upcoming strikes. The Twitter user “RJD64” presumably refers to the possibility that these protective measures could somehow break down, leading to direct conflict between the US and Russia and ultimately nuclear war.
Nassetta and Fecht seemingly believe this warning can be ignored as simple Russian propaganda. What the authors fail to mention is that the very “Western institutions” they claim to defend, like the Bulletin of the Atomic Scientists, the Middle East Institute, and even Bloomberg, have made exactly the same warning, suggesting that nuclear destruction could result from a hot war with Russia, particularly at the eastern European NATO border or in Syria. After all, US-Russian tensions have brought the world alarmingly close to nuclear catastrophe on more than one occasion. Recall that Nassetta and Fecht are writing for the “Center for Nonproliferation Studies”.
A second “dismaying” tweet (Fig. 4, bottom) intimidates “average Westerners” by “suggesting that the Assad regime acts as a bulwark against jihadism in Syria [and invoking] the consequences of past US military policy failures to deter similar regime change efforts in Syria.” Specifically, the user posits that the removal of Assad could empower al-Qaeda and ISIS and create a refugee crisis like that followed the NATO-led ousting of Gaddafi in Libya. Once again, Nassetta and Fecht demonstrate their curious respect for “Western institutions” by completely ignoring the fact that RJD64 is simply repeating the opinion of a US intelligence agency, the DIA, whose aforementioned 2012 report also predicted the destabilization of the country could embolden jihadists.
Nassetta and Fecht also warn the reader not to trust the tweet’s “dismaying” comparison between Libya and a potential American regime change operation in Syria. But, why is such a comparison unwarranted? Both the 2011 American intervention in Libya and the 2003 invasion of Iraq famously destabilized their respective targets, generating terrible violence, sectarian strife and severe refugee crises[iv]. Barack Obama, leader of that important “Western institution” called the US military, even referred to the American intervention in Libya as the greatest mistake of his presidency. Nassetta and Fecht seem to think that Syria is somehow different from these previous engagements, but they provide no justification.
Apparently, Russian agents are less adept at this tactic, because Nassetta and Fecht only provide one example, from Twitter user “Jella Cole” (Fig. 5.). Cole claims without evidence that the Syrian war is staged by the CIA and that Assad had no reason to use chemical weapons. While it is true that Assad would have been unwise to use chemical weapons when he didn’t need to, this is not proof of his innocence. For example, Saddam Hussein probably did not need to gas thousands of Iranians during the Iran-Iraq War using chemicals given to him by the United States, but he did anyway.
Less interesting than this sole example of a “dismiss” tweet are some of the general themes that characterize similar tweets, according to Nassetta and Fecht. For instance, the authors claim that Russian agents use the dismissal tactic to argue “[t]he West is not interested in verifying these allegations [of chemical weapons use] as it was waiting for a reason to launch a war.” What the authors fail to mention is that this is completely true. The US, France and Britain were patently uninterested in verifying the allegations against Assad, since they bombed long before anyone from the OPCW could even begin to investigate the April 7th Douma incident.
The new political objectivity
Suppose that all of the accounts reviewed by Nassetta and Fecht were actually run by Vladimir Putin personally. What would Putin’s hidden propaganda message be? On the one hand, the message would consist of some scattered racist and conspiratorial musings together with pro-Trump slogans like “make American great again.” The report’s authors do not seem to care very much about this aspect of the “disinformation campaign”. Instead, the they focus on the truly diabolical intention: “to deter the use of military force by the United States.” Putin would be influencing the “genuine Western” mind with phrases like “I don’t think their [sic] is anything wrong with questioning the official narrative. The MSM [mainstream media] does lie” and “I voted for Trump because he was against the wars. Won’t be voting for him again.” and with hashtags like “#TrumpsIsANeoCon” and “#EndWar”. Perhaps “Putin’s” message is already taking hold, since Nassetta and Fecht conclude that “[a]lthough ultimately unsuccessful, their [Russian agents’] operational narratives abounded in the discourse among the political right” and “it seems reasonable to assume that at least some sectors of the domestic Western populace were affected by the discourse.”
Of course, nowhere do Nassetta and Fecht actually consider the possibility that intervention in Syria could be both a great disaster and a great crime. Nowhere do they actually consult the documentary record from journalists in the region, or from the OPCW or UN, or from that lauded “Western institution”, the Defense Intelligence Agency. If they did, they would note that the Syrian conflict is a tangle of propaganda from both sides that has gradually stabilized despite extreme violence from both the heinous Assad and his jihadist opponents. It is very much not a playground for the dangerous and failed strategy of American meddling or invasion. In that sense, calling for skepticism about the official claims made by the US government, warning about the dangers of conflict with Russia and noting the very real possibility of jihadist expansion are not distorting, distracting, dismaying or dismissing. They are just truisms, which Nassetta and Fecht would likely defend if the country in question were anything other than the United States. Or would they also trust Russia’s humanitarian justification for the 2014 annexation of Crimea?
One might argue that the Nassetta-Fecht report is not a political document per se, but simply a dispassionate, empirical analysis. The authors examined a large number of tweets and the patterns of synthetic actors naturally emerged from the data. Their policy recommendations, which include asking “high influence individuals” not to make “contentious statements” and calling on Twitter to verify the identities of normal users, are just logical conclusions, not opinions.
Yet, the whole of the Nassetta-Fecht analysis is transparently based on US foreign policy dogmas, notably the assumed justification of American military intervention. The report’s statistical veneer is merely stylistic, designed to make the authors’ premises seem absolute and objective, a common trend in contemporary data science. This places the report alongside recent efforts to use more sophisticated algorithms to detect and block “fake news”, algorithms often trained with hand-labeled “real” and “fake” news articles. While the ultimate (one could argue absurd) goal of this new political “objectivity” is to create an algorithm that can automatically distinguish truth from falsehood, current systems simply say whether or not a given news item “agrees” with “high-credibility news outlets.” Otherwise, it is declared false or suspicious.
The deployment of similar algorithms has already had a serious effect on dissident media. After Google re-calibrated its search algorithm to combat fake news in the Trump era, numerous left media organizations, like Democracy Now!, Counterpunch and the World Socialist Website, saw a huge decline in traffic. Facebook, for its part, has been intermittently blocking the pages of left-leaning organizations like Venezualanalysis.com and teleSUR English. On October 11, Facebook announced its plans to remove pages displaying “inauthentic” behavior, a purge that resulted in the deletion of around 800 pages, including those concerning anti-war activism and the monitoring of police brutality. Meanwhile, establishment media organizations like The New York Times and The Washington Post, which, not ten years ago, were regularly criticized for their thoughtless endorsement of the Iraq War, have a renewed cachet. One of the many unintended bad consequences of the Trump administration has been to make people take The Washington Post‘s new self-serving slogan seriously: “Democracy Dies in Darkness”.
Nassetta and Fecht have not only caught the attention of the mainstream press, but also the US State Department, who have apparently invited them to a meeting to discuss their findings. Such collaboration between “data-based” foreign policy analysts and the US government threatens to create a dangerous feedback loop in which “objective” algorithms guide policy that is then covered by news articles curated by the same methods. And, unlike the silly “memes” with which Russian agents supposedly undermined US democracy in 2016, matters of war and peace are serious. Foreign policy analysts should take care to remember that their recommendations refer to real people who are suffering at the hands of the United States and other great powers. One doesn’t need a data set of 850,000 tweets to see that.
René Gateaux, founder of Brown War Watch, is a PhD candidate at Brown University studying computer vision and machine learning. His pen name honors the eponymous French mathematician killed at age 25 in World War I.
[i] Cockburn, P. (2017), p. 273. The Age of Jihad: Islamic State and the Great War for the Middle East. London: Verso.
[ii] Ibid., p. 290
[iii] Incendiary white phosphorous, though technically not classified as a chemical weapon by the OPCW, cannot be used as an anti-personnel weapon, except when methods less likely to result in unnecessary suffering are unavailable, according to customary laws of war. It can never be used against civilians by Protocol III of the UN’s Convention on Certain Conventional Weapons. For its part, the US has both used white phosphorous against Iraqis while claiming it is not a chemical weapon and accused Iraq of using white phosphorous while claiming it is a chemical weapon.
[iv] See chapters 2-7 and 9-10 in reference 1, as well as the Iraq chapters from Robert Fisk’s The Great War for Civilization (2005) for good firsthand accounts.