Characterizing the Dynamics of Conspiracy Related German Telegram Conversations during COVID-19

2025-07-16

AI Summary

This study investigates the dissemination mechanisms of conspiracy-related content on Telegram in Germany during the COVID-19 pandemic and its implications for social trust, democratic resilience, and public health. Leveraging large-scale German-language chat data, the analysis integrates geolocation tagging, temporal sequence modeling, and multilayer network analysis, augmented by NewsGuard credibility assessments to classify shared URLs by source reliability. Results reveal structural concentration: only 5% of core groups disseminated 94% of conspiracy content, with 43% of shared links pointing to low-credibility sources. Information propagation follows a unidirectional, hierarchical pattern—from nationally oriented broadcast channels to locally embedded, low-activity groups—with sparse inter-group connectivity. Temporal activity peaks significantly correlate with key pandemic milestones. This work constitutes the first systematic empirical characterization of structural centralization and spatiotemporal evolution of conspiracy diffusion within Telegram’s decentralized architecture, offering evidence-based insights for platform governance and targeted risk mitigation interventions.

Technology Category

• Application Domains: Misinformation & Fake News
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust

Application Category

• Web Mining and Content Analysis: Content-based information diffusion
• Responsible Web: Mitigating misinformation and disinformation
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Conspiracy theories have long drawn public attention, but their explosive growth on platforms like Telegram during the COVID-19 pandemic raises pressing questions about their impact on societal trust, democracy, and public health. We provide a geographical, temporal and network analysis of the structure of of conspiracy-related German-language Telegram chats in a novel large-scale data set. We examine how information flows between regional user groups and influential broadcasting channels, revealing the interplay between decentralized discussions and content spread driven by a small number of key actors. Our findings reveal that conspiracy-related activity spikes during major COVID-19-related events, correlating with societal stressors and mirroring prior research on how crises amplify conspiratorial beliefs. By analysing the interplay between regional, national and transnational chats, we uncover how information flows from larger national or transnational discourse to localised, community-driven discussions. Furthermore, we find that the top 10% of chats account for 94% of all forwarded content, portraying the large influence of a few actors in disseminating information. However, these chats operate independently, with minimal interconnection between each other, primarily forwarding messages to low-traffic groups. Notably, 43% of links shared in the data set point to untrustworthy sources as identified by NewsGuard, a proportion far exceeding their share on other platforms and in other discourse contexts, underscoring the role of conspiracy-related discussions on Telegram as vector for the spread of misinformation.

Problem

• Analyzing German Telegram conspiracy chats during COVID-19
• Mapping information flow between regional and influential channels
• Assessing misinformation spread via untrustworthy source links

Innovation

• Geographical, temporal, and network analysis of Telegram chats
• Examining information flow between regional and influential channels
• Identifying top chats' disproportionate influence on content spread