Malicious earworms and useful memes, how the far-right surfs on TikTok audio trends

2025-06-25

AI Summary

This study examines how far-right actors exploited TikTok’s audio trend infrastructure to embed xenophobic content within innocuous memes during the 2024 German state elections, thereby evading content moderation. Employing digital ethnography and cross-meme comparative analysis, the authors reverse-trace audio propagation pathways and variant evolution. Findings reveal that TikTok’s “sound infrastructure” constitutes a critical moderation blind spot: hate-laden audio—though demoted in recommendation algorithms—remains discoverable via associated benign hashtags in search results, enabling covert, persistent online visibility. The study makes three key contributions: first, it systematically demonstrates how audio’s inherent ambiguity, mass appeal, and infrastructural embedding jointly amplify extremist dissemination; second, it theorizes the “acoustic covert diffusion” mechanism; and third, it provides empirically grounded insights and a novel theoretical framework for platform governance and counter-extremism policy.

Technology Category

• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Application Domains: Humanities & Computational Social Science
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Web Mining and Content Analysis: Content-based information diffusion
• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization
• Security and Privacy: Data transparency and provenance

Abstract

With its features of remix, TikTok is the designated platform for meme-making and dissemination. Creative combinations of video, emoji, and filters allow for an endless stream of memes and trends animated by sound. The platform has focused its moderation on upholding physical safety, hence investing in the detection of harmful challenges. In response to the DSA, TikTok implemented opt-outs for personalized feeds and features allowing users to report illegal content. At the same time, the platform remains subject to scrutiny. Centering on the role of sound and its intersections with ambiguous memes, the presented research probed right-wing extremist formations relating to the 2024 German state elections. The analysis evidences how the TikTok sound infrastructure affords a sustained presence of xenophobic content, often cloaked through vernacular modes of communication. These cloaking practices benefit from a sound infrastructure that affords the ongoing posting of user-generated sounds that instantly spread through the use-this-sound button. Importantly, these sounds are often not clearly recognizable as networkers of extremist content. Songs that do contain hateful lyrics are not eligible for personalized feeds, however, they remain online where they profit from intersecting with benign meme trends, rendering them visible in search results.

Problem

• Examining far-right exploitation of TikTok audio trends for extremist content
• Analyzing how xenophobic messages hide in benign meme trends
• Assessing TikTok's moderation gaps in detecting cloaked hate speech

Innovation

• TikTok sound infrastructure enables extremist content cloaking
• Moderation focuses on harmful challenges detection
• Opt-outs and reporting features comply with DSA