06

Disinformation on messenger services worldwide: A study of Germany

German case study

Infrastructure

To understand and compare both the effects and the spread of disinformation through messenger services worldwide, it is necessary to consider the broadband and mobile coverage of individual countries, as well as economic factors such as the cost of internet. Germany‘s mobile coverage and 4G expansion is mediocre, especially in rural areas19, and the mobile data prices are amongst the highest in Europe20. However, when the average net income is taken into account the costs are low, especially by international standards. Almost 50 percent of Germans pay less than 30 EUR and 23.08 percent pay less than 20 EUR a month for mobile contracts21. With an average net household income of 3,661 EUR22, 20 EUR is equivalent to about 0.5 percent of net household income. Smartphone market saturation in Germany is only at around 79 percent23.

News consumption

Germans primarily get their news from television. At 72 percent, TV is ahead of online sources – including social media – which makes up 68 percent. Over six years, the consumption of news in print media has almost halved, from 63 percent in 2013 to 34 percent in 2019. Social media as a first choice for news increased from 18 to 34 percent over the same period. The use of smartphones as the primary device for accessing news is steadily increasing but at 55 percent is still behind the computer, which sits at 56 percent. By international standards, trust in the media is in the top third of countries. However, the Relotius case24 reduced trust by three percent on the previous year, according to the Reuters Digital News Report 2019.

Mr Relotius was a reporter for the magazine ‚Der Spiegel‘ who fabricated articles. The highest trust levels are awarded to public broadcasters such as ARD, whose ‚‘Tagesschau‘ ranks higher than the comparative ‚ZDF heute‘ programme. Regional and local news follow just behind and then national newspapers like Süddeutsche Zeitung, Frankfurter Allgemeine Zeitung, or weekly magazines like Focus and Der Spiegel.

By contrast, at only 16 percent, trust in news from social media in Germany is low. The picture is different for search engines, which 27 percent of Germans trust to provide their news. When we compare social media platforms through which news is consumed, Facebook has the largest share, with 22 percent. YouTube added four percent on the previous year and is in second place at 19 percent, with WhatsApp increasing by two percent to 16 percent. At four percent Facebook Messenger is almost irrelevant when it comes to news consumption in Germany. According to Reuters Report, 22 percent of Germans share news via social media platforms and just 14 percent comment on news25.

Disinformation in German elections

Considering the breadth of disinformation created by domestic and foreign actors during the 2016 US presidential election, there was worry in Germany about disinformation in the run-up to the 2017 federal election. Part of the German Network Enforcement Act (Netzwerkdurchsetzungsgesetz, NetzDG), which came into effect in October 2018, was initially intended to target fake news, as described in the introduction of the draft bill26. However, this term was soon removed when it became apparent that false news and disinformation are not usually illegal, rendering any possible intervention by law enforcement a moot point. A suspicion that automated accounts may be used to interfere in the election led to hearings on social bots in the German parliament. The Office of Technology Assessment for the German parliament reached the conclusion that social bots were not a threat27. Alexander Sängerlaub, Miriam Meier and Wolf-Dieter Rühl looked at whether and how disinformation may have been spread in the run-up to the 2017 federal election28. In a comprehensive study, they were able to show how misinformation and disinformation was spread via social media networks in Germany. Most were cases of mis- and disinformation where true statements were taken out of context, or statements not published in their entirety. The study showed that established media companies were also involved in the spread of mis- and disinformation. In one case, misinformation that spread rapidly originated from the German Press Agency (dpa). The dpa incorrectly used information from a police press release regarding a number of incidents at Schorndorf Volksfest festival. Based on a police report stating that 1,000 young people of mostly immigrant background had gathered in the Schlosspark, the dpa reported the following: „According to the police, up to 1,000 people gathered and rioted in the town‘s Schlosspark on Saturday night. A large number of them came from immigrant backgrounds.“29. The report was circulated through the online news channels of Stuttgarter Nachrichten, Welt and SWR, without the editors double-checking its veracity. The right-wing party AfD (Alternative für Deutschland) used the report for political purposes.

The study concluded there was no major presence of disinformation during the campaign and no pre-election disinformation campaigns. The spread of the case studies reviewed was manageable, provided that established media outlets were not involved in dissemination. The authors concluded that most mis- and disinformation centred around the topics of refugees and immigrants and was mostly picked up by the AfD. Because their voters lack trust in established media outlets, their primary source of news are social media platforms and right-wing populist media outlets such as the Epoch Times. A survey was also able to show that voters believe in mis- and disinformation when it confirms their world-view. Through extensive research, journalist Karsten Schmehl showed that manipulative campaigns on Twitter were not carried out by social bots but through coordinated action by people; so-called trolls30. During a TV election debate, several hundred people who belong to the right end of the political spectrum tweeted under the hashtags #Kanzlerduell and #Verräterduell (#ElectionDebate and #TraitorDebate respectively). This group had previously developed a number of memes on 4Chan forums with the intent of influencing the federal election. The campaign was organised and coordinated via the messenger service Discord, which included links to tweets from individual trolls on their numerous Twitter accounts. The plan was for others to like and retweet them in order to expand their reach and get the hashtag #Verräterduell to trend in Germany. When they failed to achieve their goal, a decision was made to refocus their efforts on YouTube.

Facebook posting of the AfD from 07/16/2017

Messenger services in political communication

The ‘messengerisation‘ of communication, a term coined by political advisor Martin Fuchs, is currently a marginal phenomenon in German political communication. While WhatsApp and others were used by parties in the 2017 federal election, the focus was on traditional platforms such as Facebook, Twitter and Instagram. The CDU (Christlich Demokratische Union Deutschlands) and SPD (Sozialdemokratische Partei Deutschlands) both ran dedicated channels on messenger services, which were primarily used to provide party members with the latest information. On election day, parties used WhatsApp to encourage people to vote – sometimes with pre-made messages31. There is no information that supports the notion that large-scale mis- or disinformation was shared through messenger services during the election. To date, there hasn‘t been a thorough review of how Germans use groups on messenger services. For instance, in the chapter on groups in the 2019 Reuters Digital News Report, Germany isn‘t discussed in relation to news consumption through messenger services. Neither does the country report on Germany mention groups, so we can assume that (large) groups, especially those that include people outside one‘s social circle have so far not had a measurable impact on political communication and news consumption in Germany. The increasing importance of groups on messenger services can be analysed by studying their use by right-wing radicals. The Amadeu Antonio Foundation has studied the phenomenon in detail32. Right-wing networks on messenger services, sometimes referred to as dark social, received more press attention following the right-wing terrorist attacks in Halle and Hanau. One of the largest and well-known German-speaking right-wing groups was created by the Identitarian movement figurehead Martin Sellner. After being blocked on Facebook and Instagram, the group moved to the messenger service Telegram – in part due to calls from the American Neo-Nazi website ‚Daily Stormer‘. While WhatsApp groups are limited to 256 subscribers, Telegram groups can include up to 200,000. Telegram channels can be subscribed to by an unlimited number of people. Telegram, which relocated its headquarters from Russia to Dubai, has said it will disclose IP addresses and phone numbers on court orders. Nonetheless, the messenger service currently seems the preferred option for sharing radical right-wing content – and similarly radical Islamist content – possibly because it expands the reach to people not on boards like 4Chan. Sharing radical right-wing content can thus become normalised and a part of everyday communication. According to the Amadeu Antonio Foundation, Martin Sellner‘s Telegram channel had 39,000 subscribers in September 2019. He largely uses the channel to share links to his YouTube videos, memes or third-party videos and encourages people to establish local groups for networking. Unlike WhatsApp which requires a user‘s phone number, Telegram includes a feature to show other users in the area and add them to groups. According to the study, radical right-wing groups, in particular, make use of this feature. The Amadeu Antonio Foundation analysed 197 Telegram channels and 38 radical right-wing groups. Analysis of the content showed that the channels are primarily used as news aggregators, but also as a space for the community of YouTubers. The biggest motivation for using these channels is to network with other radical right-wingers, presumably because it creates a safe space for them to exchange ideas. The radical right-wing prepper networks‘ Nord-/Ost-/Süd-/Westkreuz and ‚Revolution Chemnitz‘ planned and coordinated attacks using the channels.

19 Cf. Balser 2019. 20 Cf. Verivox 2018. 21 Cf. Statista 2018a. 22 Cf. Statistisches Bundesamt 2020. 23 Cf. Initiative D21 2020. 24 Cf. Yuhas 2018. 25 Cf. Newman et al. 2019. 26 Cf. Bundesministerium der Justiz und für Verbraucherschutz 2017. 27 Cf. Kind et al. 2017. 28 Cf. Sängerlaub, Meier, and Rühl 2018. 29 quoted in ibid. 30 Cf. Schmehl 2017. 31 Cf. Voigt and Seidenglanz 2017. 32 Cf. Dittrich et al. 2020.