TikTok Rewards Divisive Political Messaging During the 2025 German Federal Election

2025-09-12

AI Summary

This study investigates how TikTok’s algorithmic recommendation system influenced political information dissemination during the 2025 German federal election, specifically addressing whether divisive political content achieves higher user engagement. Analyzing 25,292 TikTok videos posted by German politicians, we employed computational content analysis to quantify fine-grained emotional valence (e.g., anger, disgust) and ideological orientation (in-group vs. out-group targeting). Results demonstrate that content expressing negative emotions and out-group hostility significantly increases user interaction rates; far-right and far-left parties not only produced such content more frequently than centrist parties but also attained disproportionately higher platform visibility per post. This is the first empirical study to document, within an electoral context, a structural algorithmic bias on short-video platforms favoring divisive political expression. Our findings provide critical evidence for understanding how platform architectures reinforce political polarization in digital public spheres.

Technology Category

• Humans and AI: Voting
• Search and Optimization: Algorithm Configuration
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Short-form video platforms like TikTok reshape how politicians communicate and have become important tools for electoral campaigning. Yet it remains unclear what kinds of political messages gain traction in these fast-paced, algorithmically curated environments, which are particularly popular among younger audiences. In this study, we use computational content analysis to analyze a comprehensive dataset of N=25,292 TikTok videos posted by German politicians in the run-up to the 2025 German federal election. Our empirical analysis shows that videos expressing negative emotions (e.g., anger, disgust) and outgroup animosity were significantly more likely to generate engagement than those emphasizing positive emotion, relatability, or identity. Furthermore, ideologically extreme parties (on both sides of the political spectrum) were both more likely to post this type of content and more successful in generating engagement than centrist parties. Taken together, these findings suggest that TikTok's platform dynamics systematically reward divisive over unifying political communication, thereby potentially benefiting extreme actors more inclined to capitalize on this logic.

Problem

• Analyzing political message traction on TikTok
• Examining algorithm-driven engagement with divisive content
• Investigating platform rewards for extreme political communication

Innovation

• Computational content analysis of political videos
• Measuring engagement with negative emotional content
• Comparing extremist versus centrist party performance