Generative <scp>AI</scp> for Social Interaction Analysis: A Study of <scp>COVID</scp>‐19 Video Comment Threads

Abstract This study examined the use of generative artificial intelligence, specifically large language models, to analyze social interactions on video‐sharing platforms during the COVID‐19 pandemic. Group dynamics, sentiment, and communication patterns were investigated using interaction process analyses in four studies: Study 1 clustered videos based on shared user engagement; Study 2 classified comments using interaction process analysis with open‐source large language models; Study 3 analyzed reply threads to capture interaction dynamics; and Study 4 compared classification consistency across large language models and against human annotation. The findings underscore the scalability and efficiency of generative artificial intelligence in processing large datasets, revealing its ability to uncover complex interaction patterns. Simultaneously, the study highlights key challenges, including cost differences and variability in the accuracy of models, underscoring the need for further research to refine artificial‐intelligence‐based methodologies in social psychology.