Discourse profiles improve detection of climate misinformation online
Modeling Discourse Profiles in Climate Misinformation Discourse
Computation and Language
Summary
Detecting climate misinformation online usually looks at one comment at a time, missing the bigger picture. The authors show that misinformation often follows repeated patterns in conversations, like common ways of arguing or phrasing. They grouped TikTok comments about climate into these patterns, called discourse profiles, then added this to their detection methods. This approach helps find misinformation more accurately and reveals how certain conversation styles affect how misinformation spreads and is accepted.
What this means in practice
- •For social media platform engineers: Integrate discourse profile features to enhance automated climate misinformation detection in user comments on video platforms.
- •For online content moderation teams: Use discourse profile–based tools to better identify and monitor recurring misinformation patterns for improved moderation decisions.
Tested on one dataset.
Authors
Ge Gao, Zhengyang Shan, Bryan A. Plummer, Margrit Betke, Derry Wijaya
Abstract
Prior work on climate misinformation detection typically classifies individual comments or claims using text alone. This framing overlooks a key property of online climate discourse: misinformation is often produced, reinforced, and sustained through recurring rhetorical patterns that reappear across comments and threads. Using a large-scale TikTok climate discourse dataset, we construct discourse profiles by clustering climate-related comments into recurring rhetorical patterns. We then incorporate these profiles into fine-grained climate misinformation classifiers using a late-fusion architecture. Under video-disjoint evaluation across three training seeds, discourse-profile-aware models significantly improve macro-F1 over text-only baselines for bag-of-words, TF-IDF, and MPNet representations. RoBERTa-large exhibits a positive mean improvement but substantially greater variability, and the difference is not statistically significant. Beyond classification, our results show discourse profiles shape misinformation reception and propagation. Discourse-profile-conditioned language model agents exhibit misinformation acceptance patterns aligned with profile-level misinformation prevalence in ClimateDisc, and the relationship between misinformation prevalence and engagement persistence differs across dominant discourse profiles.