Researchers have developed a new framework for analyzing the complex claims made in health-related social media videos. This framework models discourse by linking atomic claims to thematic aspects, stance, and pragmatic attributes. A benchmark dataset of 1,191 annotated claims from 60 videos was created to evaluate automated analysis. While current large language models perform well on thematic categorization and stance prediction, they struggle with detailed pragmatic profiling, suggesting a need for specialized inference strategies. AI
IMPACT This research could lead to more nuanced AI analysis of health information on social media, improving understanding of misinformation and public health communication.
RANK_REASON This is a research paper describing a new framework and benchmark for analyzing discourse in health narratives. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
