Inevitable Exposure: Analysing the Privacy Paradox in the Age of Digital Connectivity with Machine Learning Paradigms

This study investigates the privacy paradox, a phenomenon whereby individuals express concerns about privacy while actively engaging in activities that compromise their personal data. Through a topic analysis of video transcripts from Douyin, China's most popular short video-sharing platform, we investigate the application of text analytics to examine the privacy paradox phenomenon. We leverage BERTopic, a state-of-the-art topic modelling technique, and the Meta Llama 3 Large Language Model to conduct a topic analysis of discourses related to privacy issues in Douyin videos. Our analysis generates useful insights into where and how the privacy paradox manifests in various aspects of daily life. The findings underscore the need to put in place proactive data protection measures to address public concerns about privacy.