New AI Framework Tracks User Stance and Misinformation at Scale

A team of researchers from Macquarie University (Seneviratne, Ikram, Vatsalan, Asghar, Kaafar) has developed a scalable AI-driven analytics framework designed to monitor user engagement and stance detection on social media platforms. The modular pipeline combines data ingestion, filtering, topic modeling, sentiment analysis, and stance detection, all optimized for large-scale, real-world datasets. In their evaluation, the team analyzed over 7 million user comments from nearly 50,000 YouTube videos associated with conspiracy narratives. The results show that conspiracy content attracts up to 70% of total user engagement within the first week of publication, revealing strong early amplification dynamics. Additionally, the system identified a subset of highly active users who dominate engagement across multiple videos and channels.

Stance analysis further revealed that a majority of users in these communities express favorable positions toward conspiracy content, underscoring the role of user communities in reinforcing harmful narratives. The framework demonstrates the feasibility of deploying service-oriented analytics for real-time monitoring of user behavior at platform scale. This work moves beyond traditional behavioral or content-centric analyses by offering a practical, scalable solution for platforms to detect and mitigate misinformation early. The paper is available on arXiv (arXiv:2605.29199) and is 11 pages long, covering both the technical architecture and empirical findings.