Well before fake news became a thing, Karl was reporting on the fascinating details that have emerged about Russia’s Internet troll factories that relentlessly pump out fake posts on an extraordinary scale. More recently, the Russian Defense Minister Sergei Shoigu revealed that the country’s military has created a force specifically tasked with waging information warfare. We may know about Russia’s domestic activities in this area, but what about online propaganda teams active in other countries? One data point towards answering that question is provided by an article on a site called the Disinformation Review, which describes itself as follows:
the latest cases of news articles carrying key examples of how pro-Kremlin disinformation finds its way in international media, as well as news and analysis on the topic. The review focuses on key messages carried in international media which have been identified as providing a partial, distorted or false view or interpretation and/or spreading key pro-Kremlin messaging.
It does not necessarily imply however that the outlet concerned is linked to the Kremlin or pro-Kremlin, or that it has intentionally sought to disinform. The Review is a compilation of cases from the East StratCom Task Force’s wide network of contributors and therefore cannot be considered an official EU position.
That is, the Disinformation Review draws on information provided by the EU-funded East StratCom Task Force, and is part of the EU’s response to what it sees as growing Russian propaganda directed against the European Union and its member states. A recent post on the site delves into another troll factory, but this time in Sweden. It reports on an article originally published by the Swedish daily Eskilstuna Kuriren:
we read that Swedish trolls primarily target journalists; that they develop and use scripts for their telephone conversations; and that the trolls are paid 1,000 SEK (110 EUR [about $110]) when their recorded telephone conversations obtains enough ‘likes’ in social media. We read that the trolls work with manuals that instruct them to edit the recordings to make them as “entertaining” as possible. We also read that the people behind the troll factory belong to Swedish racist and extreme right wing organisations.
But it’s not only extreme right-wing viewpoints that the Swedish Internet troll factory supports:
The agenda of the political movement affiliated with the trolls is, according to the investigation, “xenophobia and Islamophobia”, combined with promotion of commentators who “support Russia after the occupation of the Crimea and the Russian-backed civil war in Ukraine”.
Despite that intriguing fact, the Swedish newspaper report was unable to establish who was funding the propaganda efforts. However, it does provide some interesting information about what makes a successful Internet troll factory:
Eskilstuna Kuriren ends their piece by asking that question to Jack Werner, co-founder of the popular Swedish fact-checker Viralgranskaren. According to Werner, the organisation is possibly limited in size, but a central part of its strategy is to make itself look very big: “The aim of propaganda is to respond to light so as to make the shadow it casts is as large as possible. If you really want to give the impression that your side is the largest, most dedicated and most passionate, it requires more work, for example, you will need to spend days and nights writing comments in the internet.”
The article quotes figures from an earlier investigation into right-wing propaganda sites, which found that just 183 individual writers accounted for 366,291 comments out of a total of half a million, which works out as 2,000 comments per person on average. Perhaps that high volume could be turned against the Internet troll factories.
Since it is very hard for people to write so many comments so quickly without using similar phrases, sites might check new posts against old ones to eliminate those that are likely to be from a few writers churning them out to order. The technology already exists, and is widely used to spot academic plagiarism, for example. Cloud computing platforms would allow this approach to be applied routinely at a reasonable cost, and there would be scope for new third-party services to flag up re-used content across multiple sites. Google’s parent company, Alphabet, is already working on software in this area. Maybe the time has come to apply a little more intelligence and computational firepower to tackling the growing threat that intentionally misleading and inflammatory posts by Internet trolls represent not just to online discourse, but far beyond.