Topical collection hate speech
Why digital behavioural data on hate speech?
Hate speech on the internet and on social media has increasingly moved into the focus of public and academic attention in recent years. Digital platforms enable extremist actors—particularly from the right-wing spectrum—to deliberately bypass established gatekeepers of political discourse such as political parties or mainstream media. This allows them to disseminate their ideologies, narratives, and enemy images in an unfiltered and often anonymous manner. This form of digital communication poses a serious challenge to democratic societies, as it is explicitly designed to undermine trust in democratic institutions and to deepen social divisions. The German right-wing extremist scene, too, has recognised the mobilising potential of social media and makes use of a network of different platforms as well as high-reach accounts of politicians and influencers. Increasingly, this also includes content generated with generative AI, which is used in particular to appeal to younger target groups (Boyd, 2010; Winter, 2019; Rothut et al., 2023). Adolescents, who are in a sensitive phase of identity and opinion formation, are especially susceptible to such manipulative content (Gaden 2014).
Although progress has been made in recent years in providing German-language datasets on digital hate speech (https://hatespeechdata.com/), the number of reusable, high-quality, and well-documented corpus datasets remains limited compared to the English-speaking research landscape (e.g. via SOMAR). This gap considerably hinders scholarly engagement with these phenomena and restricts the reproducibility of research findings. In order to strengthen research on extremist communication based on digital behavioural data in Germany, DP-R|EX aims to provide data on digital hate speech for reuse, while simultaneously testing, documenting, and further developing methods for data collection and data provision.
Mainstreaming of right-wing extremist and harmful content via popular online platforms
Social media platforms such as Facebook, YouTube, X (formerly Twitter), and TikTok now act as key catalysts for the mainstreaming of extremist narratives. In this context, mainstreaming refers to the process through which content that was originally marginal or extreme—often described as borderline content—gradually enters broader public visibility, gains acceptance, and ultimately comes to be perceived as normal. The boundary between what is considered legitimate and what is deemed extreme is defined by the prevailing societal discourse, which itself is a contingent and dynamic construct. This discourse actively shapes which forms of content are regarded as acceptable at a given point in time (Brown et al., 2023). Platform algorithms deliberately amplify content produced by extremist groups by integrating it into personalised feeds, thereby circumventing traditional gatekeepers of political discourse (Conway et al., 2019). At least equally important, however, are the individual choices made by users. The deliberate sharing of content, subscribing to channels, and following specific extremist actors can lead to the formation of echo chambers that reinforce pre-existing worldviews (Karell et al., 2023). In addition, extremist actors employ complex cross-platform strategies—such as cross-posting between Telegram, YouTube, and Instagram—to introduce their messages into previously non-suspect user segments (Urman & Katz, 2022; Kakavand, 2024). Through these processes of online radicalisation, hate speech and harmful content increasingly penetrate the mainstream, become visible to broader audiences, and produce offline effects, which in extreme cases may manifest in acts of politically motivated violence (Karell et al., 2023).
Telegram as a mouthpiece for racist and extremist actors
In recent years, Telegram has emerged as a central platform for racist and right-wing extremist groups, including in Germany (Zehring & Domahidi, 2023). Due to its end-to-end encryption, large channel and group capacities, and the possibility of operating anonymous broadcast channels, Telegram largely evades the oversight of formal moderation mechanisms (Dietze, 2023). Extremist actors exploit this infrastructure to disseminate propaganda videos, coded language, and memes embedding radicalised ideologies and hate speech through underground networks that are in part generated or amplified by bots and automated systems (Bloom et al., 2019). In contrast to mainstream social media platforms, Telegram does not rely on algorithmically curated feeds. Instead, content spreads virally via forwarding practices, invitation links, and cross-platform posting across thematically connected groups and channels (Walther & McCoy, 2021). As a result, hate actors are able not only to reach individuals who are already sympathetic to the extremist milieu, but also—through dense network structures—to access new audiences. This enables them to advance processes of opinion formation and radicalisation largely shielded from external intervention or regulatory oversight (Schulze, 2024).