Disinformation in the 2021 German Federal Elections: What Did and Did Not Occur

Author

Project Director "Strengthening the Digital Public Sphere | Policy", Stiftung Neue Verantwortung

The 2021 German federal election season was expected to be highly digitalized: the Covid-19 pandemic still casts a shadow over in-person events and meetings. Most political parties have held their conventions in digital or hybrid formats. In addition, parties’ reliance on online campaigning through apps, ads on websites and social media platforms has continued to grow over the past several years, since the previous parliamentary vote in 2017. Along with the expected digital campaigning came fears of higher (and worsened) online disinformation and hate speech.

These expectations and fears materialized, but only to a certain extent: it was not an all-digital campaign and there was no singular, big disinformation narrative. This, however, cannot hide the fact that Germany still lacks a fitting approach to tackle issues associated with digital campaigning and disinformation.

Big online campaigns, but putting up posters still matters

The elections were certainly highly "digital", but not quite "digital-first". It has become clear that nearly every major party, every candidate and every political campaign uses online tools to promote their messages. That is also what happened during the 2021 campaign: parties spent hundreds of thousands of euros on Facebook/Instagram and Google/YouTube alone, according to the tech companies’ own reports on political advertising. Political parties used apps to mobilize volunteers and candidates took part in online events and discussions.

However, it also became clear that campaign posters on the streets and on public transportation, as well as rallies and personal meet-and-greets remain essential to political campaigning in Germany. Candidates went out to give speeches despite the pandemic. Every few steps, candidates smiled down at pedestrians from posters on lamp posts. This means that "analog" campaigning must not be ignored yet, and that also goes for analog disinformation and hate speech. For instance, there were discussions about posters by a tiny far-right extremist party calling for Green candidates to be hung. This was supposedly tongue-in-cheek, an assertion that was roundly rejected by many people who saw the posters. Though these types of threats were rare in political advertising, negative campaigning still occurred, both online and offline. Negative campaigning refers to the denigration of one’s opponents to make oneself look good. Various candidates and parties were affected by this.

Blurred lines between negative campaigning and disinformation

Negative campaigning does not necessarily have to include disinformation at all, but it certainly can. For instance, an ad campaign against the Green candidate Annalena Baerbock, who ran on major German news websites, contained not only criticism of her and the party’s program but also mischaracterizations of her proposed policies. This campaign offers a glimpse of the blurred lines between negative campaigning and disinformation. In many cases, such messaging is legal and has been part of campaigning for decades.

Apart from the "Alternative for Germany" (AfD), all major German parties committed to a code of conduct for fair digital campaigning.

It is largely the responsibility of political parties to address this and to ensure a fair and honest campaign season. Apart from the "Alternative for Germany" (AfD), all major German parties committed to a code of conduct for fair digital campaigning. This was in part due to pressure by initiatives such as Campaign Watch, an ad hoc coalition of civil society and academic organizations calling on parties to develop a code of conduct (disclosure: the author’s employer was part of the Campaign Watch group).

Among other things, Campaign Watch suggested that parties refrain from spreading disinformation in their campaigns and that they actively counter digital violence on their social media pages. Some of the ideas were taken up by the parties, even though major gaps remained regarding restrictions of targeted political ads, for example. Such ads can flood social media and can have opaque messages tailored to specific audiences, which might distort political debates. Nevertheless, it was the first time that almost all major German parties even developed a code of conduct.

The dangers of misleading the public about the voting process

All major candidates and parties were subject to hateful commentary and, at times, online disinformation as well, as documented by German fact checkers and journalists. Contrary to 2017, when a study found that most disinformation centered around topics of migration and refugees, disinformation in the 2021 campaign was disparate and covered a variety of candidates and topics. The government’s reaction to Covid-19 was one topic that stuck out. Covid-related measures spawned protests and inspired online groups. Some of the discussions leading to these protests were laced with disinformation and contributed to an ill-informed radicalization. This bled over into campaign-related disinformation, when candidates were attacked for doing too much or too little regarding the pandemic, for instance.

Instead of a focus on a particular topic, it is interesting to note that there were instances of disinformation regarding the voting process. This had not been as prominent an issue in the 2017 campaign. Early on, the AfD was most adamant among the parties in suggesting that the mail-in ballot process was inherently unsafe, and that the German election was being manipulated. This assertion, spread through a Facebook campaign and also in speeches, wildly inflated valid concerns about mail-in ballots.

There has been no evidence of widespread, systematic voter fraud or manipulation via mail-in ballots in Germany. It is unlikely that the voter fraud narrative, reminiscent of some Republican-led messaging in the US, occurred among large parts of the German population. But it was nonetheless discussed in messaging groups and eventually made its way into official AfD online advertising.

A radicalized, violent belief that elections are fraudulent by definition might put poll workers and campaigners at risk.

Even if only small pockets of the population interacted with these lies, there is a risk that this undermines trust in the electoral process based on misleading or false information. More tangibly, a radicalized, violent belief that elections are fraudulent by definition might put poll workers and campaigners at risk (fortunately, there were no immediate reports of violence against poll workers on election day). To counter this, for the first time, the federal electoral management body was tasked with providing factual information about the voting process. Usually, this body is only concerned with organizing the voting process itself. During this campaign season, however, it started to act as a fact checker on voting disinformation.

Who was a frequent target for disinformation?

According to a representative survey in Germany, 74 percent of young female respondents perceive hate speech (very) frequently, while the average for all ages and genders was 45 percent. For disinformation, the numbers are 77 and 63 percent, respectively. With the caveat that this is a self-perception survey, it seems to be mostly women who suffer the negative consequences of hate speech and disinformation. This can lead women to refrain from speaking out online. In the German election campaign, media discussions took up the topic of gender-based disinformation and hate speech as well. The Greens reported that their candidate, Annalena Baerbock, the youngest and only female person among the declared candidates for German chancellor, was subject to online vitriol after announcing she would run. Journalists and activists provided some evidence for this. While the methodology behind some of the empirical findings was questioned, anti-hate speech groups such as HateAid as well as disinformation researchers in the US have nonetheless made compelling arguments for how these issues disproportionately affect women and minorities and fortify existing, potentially discriminatory, power structures.

A debate about sexist, anti-migrant and racist disinformation is long overdue in Germany.

A debate about sexist, anti-migrant and racist disinformation is long overdue in Germany, given previous findings in that disinformation during the 2017 campaign showed clear anti-immigrant slants. In fact, methodological questions about disinformation studies might serve to highlight the need for even better research.

This also needs to be viewed against the backdrop of the difficulties that researchers continue to face in gaining suitable data access to large platforms, especially Facebook. This problem came to the fore once again just as the election campaign was heating up, when Facebook’s actions effectively halted a German research project aiming to study campaigning on Instagram.

Lessons from the 2021 German election campaign

There are several lessons to be drawn from this first, cursory look at disinformation during the 2021 German election campaign. First of all, it is clear that binding rules for digital campaigning are needed. Relying on political parties’ and platforms’ self-commitments is not enough. While all major parties did commit to fair campaigns, a common, cross-party code of conduct was still missing and there were no clear restrictions or transparency rules regarding political ad targeting. Platforms’ voluntary transparency efforts - for instance, their political ad databases - remain flawed as well. More stringent campaign finance rules for parties and meaningful platform regulation are necessary. For the latter, the European Union’s proposed Digital Services Act and legislation on political ads transparency seem most promising. In this discussion, questions about platform design should be embraced. Instead of trying to delete disinformation (a legally and technologically futile move), it is worthwhile to consider if there can be sensible changes in platform design to slow the spread of disinformation.

Second, the false narrative of voter fraud via mail-in ballots shows that disinformation is not just relegated to personal or policy-related topics, but also to the voting process itself. It also shows that disinformation playbooks from elsewhere can be applied to Germany. After the study on disinformation during the 2017 election campaign showed that it was mostly domestic and not foreign actors spreading disinformation, the debate was broadened beyond foreign interference. Similarly, now, it is necessary to widen the scope from policy-related disinformation to disinformation related to political processes such as (mail-in) voting.

Third, contrary to 2017, there were many more researchers, activists and organizations covering disinformation during the elections. The topic was still not highest on the agenda, but surveys, campaigns, research websites and media coverage at least contributed to a steadier discussion and raised awareness of the potential threats. The flipside of this positive development is the risk of people either becoming numb to the phenomenon or being more scared by the reports about disinformation than the actual disinformation out there. This supports the pleas that academic and civil society experts have made for years to also consider the role of traditional media in amplifying (online) disinformation narratives. "De-sensationalizing" studies and reports about disinformation can be one way to improve research and reduce potential harms from disinformation.

Lastly, policymakers should not wait for a single, all-consuming disinformation narrative to disrupt an election to act, but rather realize that various parallel disinformation narratives and groups cannot be ignored as a potential factor in shaping political debates. To that end, countermeasures need to evolve and be better coordinated. The ingredients for Germany’s recipe to tackle disinformation remained largely the same, but with increased doses: a little more fact-checking, an extra heap of research and studies, some more institutional protections regarding voting disinformation, and some added media attention for flavor. In the future, it might be necessary to consider the mix of ingredients a little more thoroughly and add new ingredients such as campaign and platform regulation.

Copyright: John MACDOUGALL / AFP