Dynamics of Algorithmic Content Amplification on TikTok

2025-03-26

AI Summary

This study investigates how TikTok’s recommendation algorithm dynamically amplifies interest-aligned content in response to user viewing behavior and its impact on content diversity. Employing a multi-interest sock-puppet auditing framework, the authors integrate time-series analysis, Markov modeling, and cross-interest controlled experiments. Their key contributions are: (1) empirical identification of a two-phase dynamical pattern—“reinforcement phase” followed by “diversity decay phase”—where strong interest feedback loops emerge within just 200 views; (2) discovery of topic-specific amplification bias, with significant variation in amplification intensity across interest domains; and (3) demonstration of a strong negative correlation between amplification intensity and user exploratory behavior, leading to progressive diversity erosion and declining exposure to novel topics. These findings reveal early-stage mechanisms underlying interest ossification and diversity suppression in short-video recommender systems.

Technology Category

• Data Mining & Knowledge Management: Recommender Systems
• Machine Learning: Learning Preferences or Rankings
• Application Domains: Humanities & Computational Social Science

Application Category

• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems
• Web Mining and Content Analysis: Content-based information diffusion
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Intelligent algorithms increasingly shape the content we encounter and engage with online. TikTok's For You feed exemplifies extreme algorithm-driven curation, tailoring the stream of video content almost exclusively based on users' explicit and implicit interactions with the platform. Despite growing attention, the dynamics of content amplification on TikTok remain largely unquantified. How quickly, and to what extent, does TikTok's algorithm amplify content aligned with users' interests? To address these questions, we conduct a sock-puppet audit, deploying bots with different interests to engage with TikTok's"For You"feed. Our findings reveal that content aligned with the bots' interests undergoes strong amplification, with rapid reinforcement typically occurring within the first 200 videos watched. While amplification is consistently observed across all interests, its intensity varies by interest, indicating the emergence of topic-specific biases. Time series analyses and Markov models uncover distinct phases of recommendation dynamics, including persistent content reinforcement and a gradual decline in content diversity over time. Although TikTok's algorithm preserves some content diversity, we find a strong negative correlation between amplification and exploration: as the amplification of interest-aligned content increases, engagement with unseen hashtags declines. These findings contribute to discussions on socio-algorithmic feedback loops in the digital age and the trade-offs between personalization and content diversity.

Problem

• Quantify TikTok's algorithm-driven content amplification dynamics
• Assess speed and extent of interest-aligned content reinforcement
• Explore trade-offs between personalization and content diversity

Innovation

• Sock-puppet audit with interest-based bots
• Time series and Markov model analysis
• Quantifies amplification vs. content diversity trade-off