Doppelgänger Debunking: Monitoring Social Media Does Not Equate to Recruitment

As noted, I plan to do a more substantive piece on DOJ’s effort to disrupt Russian efforts to influence the election, but first want to debunk a few claims people are making about last week’s releases.

In this post, I debunked the claim that Lauren Chen is likely to have been targeted under FISA; FBI wouldn’t have needed FISA, when criminal process is easier to get.

There’s an even bigger error regarding something about the Doppelgänger materials released last week, traceable in significant part to this post and the screen cap from it, disseminated by others:

The screen cap comes from this passage of the affidavit supporting the take-down of a bunch of sites used by Russia’s Doppelgänger project. Gilbert and others have screen-capped primarily the part describing influencers (italicized below), without the part that directly followed, describing that Russia has a similar list of people who don’t support Russia, much less the part (bold below) describing that these were accounts were monitored to track public opinion.

66. SDA documents further reveal that SDA extensively monitors and collects information about a large number of media organizations and social media influencers. One document revealed a list of more than 2,800 people on various social media platforms like Twitter, Facebook and Telegram, spanning 81 countries, that SDA identified as influencers, including television and radio hosts, politicians, bloggers, journalists, businessmen, professors, think-tank analysts, veterans, professors, and comedians. When referring to politicians, the list often mentioned which U.S. state and/or political party they represent and the position they hold in Congress. The U.S.-based influencers accounted for approximately 21% of the accounts being monitored by SDA. On another list of over 1,900 “anti-influencers”14 from 52 countries, the U.S.- based accounts comprised 26% of the total accounts being monitored by SDA. I assess that “anti-influencer” indicates that the account posts content that SDA views as contrary to Russian objectives. Based on my review of other records obtained during this investigation, I know that SDA adds information captured through its monitoring efforts to dashboards. These dashboards analyze trends in public opinion and thereby measure the effectiveness of the malign foreign influence campaign based on its impact on public opinion. SDA’s content varies from project to project; however, it can include videos, memes, cartoons, social media posts, and/or articles. SDA’s content delivery also varies each campaign, but often relies heavily on social media posts driving targeted audiences to domains SDA controls, like the SUBJECT DOMAINS. [my emphasis]

In his story on the releases, Gilbert extrapolates from a different document that primarily focuses on using targeted advertising to attract social media users to Russian-made content, to suggest this list of 2,800 influencers might constitute those envisioned in a small section of the document as “collaborators,” though that section of the document doesn’t use the term, “collaborators.”

According to the Good Old USA project document, the Kremlin was seeking to work with influencers who are “proponents of traditional values, who stand up for ending the war in Ukraine and peaceful relations between the US and Russia, and who are ready to get involved in the promotion of the project narratives.”

Among the types of influencers listed as possible collaborators are actors, politicians, media representatives, activists, and clergymen.

The affidavit references one document maintained by the Social Design Agency, which is not included in the unsealed court documents, that contains a list of more than 2,800 people identified as influencers. While this list is global, US-based influencers account for around 20 percent of the accounts being monitored, including many US lawmakers, according to an analysis of the list by the FBI.

That is, in my opinion, a wild misreading of the material, not least because the document envisioning “working with influencers” includes passive ways to exploit pro-Russian voices, including the “rollout of real comments” from them.

Other even more inflammatory tweets have highlighted the same passage to claim that Russia is paying 2,800 people.

While it’s not clear that the FBI knows precisely what the Social Design Agency is doing with these lists, all it claims that they’re going is tracking these accounts — both pro and anti-Russian social media accounts — to “analyze trends in public opinion and thereby measure the effectiveness of the malign foreign influence campaign based on its impact on public opinion.”

There’s no claim the 2,800 people on the list are being paid.

Even if SDA were doing more, it would in no way signal full “collaboration.” An earlier report on Doppelgänger’s work (one I’m still looking for, to link), for example, described how Doppelgänger would exploit the way Elon Musk uses his Xitter account to piggyback on his visibility to magnify pro-Russian content with no involvement from him. Elmo is so predictable and so stupid with his Xitter account it requires no payment or even witting involvement to be exploited in such a way.

Similarly, there are any number of right wing members of Congress who oppose Ukrainian funding in significant part because Trump told them to; while some of them might be on the Russian payroll, the overwhelming majority are not, but they nevertheless produce social media content that is of enormous use to Russia. JK Rowling’s transphobic content similarly attracts the kind of engagement that could be usefully exploited for Russia.

The inclusion of anti-influencers on this list is a big tell that those on the influencer list are not all recruited. Indeed, my own Xitter account could be big enough and — because Musk has forced a virtual blue check on my account, increasing my visibility in algorithms — to be included on an anti-influencer account; Asha Rangappa, Tom Nichols, and Anne Applebaum are all people with credentialed anti-Russian views with more Xitter followers than me who are even more likely candidates. It often happens that trolls with their own blue checks will attempt to hijack my timeline to stir up fights; it takes aggressive blocking to prevent it.

In other words, it doesn’t take recruitment to exploit readily apparent algorithmic patterns. Even overt opposition can be harnessed, if such efforts are not aggressively combatted.

And there’s nothing in the affidavit, describing an effort to monitor public opinion, to suggest Russia is doing even that.