A Survey of Information Disorder on Video-Sharing Platforms

TL;DR

This survey addresses ID on video-sharing platforms by applying Wardle's taxonomy to map the literature across three dimensions: types of information disorder, methodological approaches, and platform features. It integrates quantitative, qualitative, and mixed-method studies, highlighting how content, engagement, and algorithmic features interact to store and amplify misinformation in video ecosystems. Key contributions include a structured ID typology application to VSPs, a synthesis of datasets and methodological trends, and a critical discussion of gaps—especially in short-form video research and cross-platform comparability. The work provides a holistic framework to guide future detection, mitigation, and sociotechnical analyses of ID on multimodal, algorithmically curated video platforms.

Abstract

Video sharing platforms (VSPs) have become central information hubs but also facilitate the spread of information disorder, from misleading narratives to fabricated content. This survey synthesizes research on VSPs' multimedia ecosystems across three dimensions: (1) types of information disorder, (2) methodological approaches, and (3) platform features. We conclude by identifying key challenges and open questions for future research.

Figures (4)

• Figure 1: PRISMA diagram of literature identification, screening, and inclusion
• Figure 2: Distribution of information disorder types across VSPs
• Figure 3: The architecture of SOTA models classifying misinformation videos
• Figure 4: Comparison of F1 by SOTA multi-modal models