January 24, 2018 / by emptywheel

 

The Gizmo™: Correlation Doesn’t Equal Adversary Nation

For days, reporters have been mis-using The Gizmo™ (the name I use for the “disinformation dashboard” from the German Marshall Fund, a black box that purports to show “Russian propaganda efforts on Twitter in near-real time”) to claim that Russian-linked accounts are pushing the #ReleaseTheMemo campaign calling for the public release of Devin Nunes’ politicized memo attacking the FBI.

As the effort lead by some Republicans to curtail special counsel Robert S. Mueller III’s investigation into the election meddling has heated up, Russian-linked accounts helped amplify a Twitter hashtag calling for the release of a memo the group hopes will help discredit Mueller’s work, according to Hamilton 68, a research firm that tracks the malicious accounts. The #releasethememo hashtag was tweeted by these accounts nearly 4,000 times in the last couple of days, the firm said.

As always with such reporting, the articles don’t provide even the nuance the project’s most responsible contributor, JM Berger, lays out on their methodology page.

  1. Not all content in this network is “created” by Russia. A significant amount—probably a majority—of content is created by third parties and then amplified by the network because it is relevant to Russian messaging themes.
  2. Not all content amplified by this network is pro-Russian. The network frequently mobilizes to criticize or attack individuals or news reports that it wishes to discredit.
  3. Because of the two points above, we emphasize it is NOT CORRECT to describe sites linked by this network as Russian propaganda sites. We are not claiming that content producers linked by this network are Russian propaganda sites. Rather, content linked by this network is RELEVANT to Russian messaging themes.

Such reports certainly don’t consider the validity of drawing conclusions from such analysis that the authors have refused to have vetted by a third party. What does it mean to openly profess to be pro-Russian, for example? Do non-consensus views on Syria or Ukraine count? Does skepticism about Russian involvement in the election count?

And the reports don’t note the serial false positives, such as the time Jim Lankford used The Gizmo™ to claim Russia was stoking tensions around NFL players taking a knee during the anthem. More responsible analysis showed that,

[B]oth #TakeAKnee and #BoycottNFL were genuinely viral movements, generating high volumes of traffic from large numbers of accounts, but both received an additional boost from bots.

The bots which amplified #TakeAKnee were primarily non-political; they appear to be bots for hire, repurposed to amplify specific posts. Of these, the most significant group is that which retweeted @DianneLogic, given its previous use in online harassment campaigns in the context of Russia and the far right. However, the evidence of its prior behavior is suggestive but not conclusive. It cannot be taken as proving Senator Lankford’s claim.

The accounts which amplifed #BoycottNFL are a different breed. They are largely cyborgs, rather than bots, posting authored content in between slews of retweets. They are also political, rather than commercial. Their sole purpose appears to be boosting far-right American posts.

In both cases, the bots were functionally anonymous, providing no verifiable information on the identity of the user behind them. There is thus no independent information which would allow us to say definitively whether they were American, linked somehow to Russia, or managed from another country entirely.

In short, in spite of this thing being shown to measure something entirely different from what reporters continue to report — correlated traffic (and that, based on unpublished criteria) rather than causal traffic — nevertheless Russia got credit for a campaign clearly driven by right wing Americans backed by a far more extensive propaganda infrastructure.

And then, even as Twitter started leaking initial analysis saying just that — that Russia wasn’t to blame …

[A] knowledgeable source says that Twitter’s internal analysis has thus far found that authentic American accounts, and not Russian imposters or automated bots, are driving #ReleaseTheMemo. There are no preliminary indications that the Twitter activity either driving the hashtag or engaging with it is either predominantly Russian.

In short, according to this source, who would not speak to The Daily Beast for attribution, the retweets are coming from inside the country.

… Two members of Congress from California, Adam Schiff and Dianne Feinstein, called on two California companies, Twitter and Facebook, to confess further manipulation by Russia.

We understand Facebook and Twitter have developed significant expertise in identifying inauthentic and malicious accounts.  Further, your forensic investigations into Russian government exploitation of your platforms during the 2016 U.S. election have helped expose to the American public the vast extent of Russia’s covert influence efforts. We therefore request that your companies conduct an in-depth forensic examination of this real-time activity on your platforms to determine:

  1. Whether and how many accounts linked to Russian influence operations are involved in this campaign;
  2. The frequency and volume of their postings on this topic; and
  3. How many legitimate Twitter and Facebook account holders have been exposed to this campaign.

Given the urgency of this matter, we ask that you provide a public report to Congress and the American public by January 26, 2018.  In addition, we urge your companies to immediately take necessary steps to expose and deactivate accounts involved in this influence operation that violate your respective user policies.

Nothing in this letter explains why Facebook should have to do this work, as The Gizmo™, the sole piece of evidence Schiff and Feinstein rely on, doesn’t track Facebook.

But even the demand to Twitter was based on yet another misreading of what The Gizmo™ actually measures. And, having never asked The Gizmo™ to explain the methodology behind its serial panics, a Senator representing both Facebook and Twitter demanded that they check its work, rather than vice versa.

If I were a forewoman in a Russian troll factory, there would be no easier way to boost my career prospects than to use a few of my bots to manipulate The Gizmo™’s sloppy methodology to claim credit for an obviously American-generated hoax. “Ивана! Давайте претендовать на последнюю республиканскую пропаганду!” Doing so would set off a self-fulfilling prophecy, precisely the kind of thing The Gizmo™’s authors claim to want to prevent, boosting Russia’s ability to sow discord with virtually no effort.

Copyright © 2018 emptywheel. All rights reserved.
Originally Posted @ https://www.emptywheel.net/2018/01/24/the-gizmo-correlation-doesnt-equal-adversary-nation/