Exploring public perceptions of generative AI and education: topic modelling of YouTube comments in Korea

Abstract

ABSTRACT This study aims to investigate the public’s perceptions regarding the integration of Generative AI (GenAI) in education by analysing comments on YouTube news clips. The study collected public comments from YouTube news clips disseminated by three prominent broadcasters in South Korea between December 2022 and June 2023. Two dimensions of public perceptions were examined: sentiments and prevalent topics. Employing machine learning techniques, we conducted sentiment analysis and topic modelling on the crowdsourced dataset of 18,566 comments from 66 YouTube news clips. The first research question focused on public sentiments towards GenAI and education. Findings reveal a predominance of neutral sentiments. Rather than adopting extreme positions of complete acceptance or rejection, the public displayed an inclination to appreciate the intricate nuances of GenAI’s implications. The second research question sought to identify the main topics emerging from public comments on GenAI and education. We identified 11 distinct topics where two topics are directly linked to educational implications: demands for changes in learning and assessment methods, and the use of GenAI in higher education. Based on the key findings, we draw implications that can inform a broader understanding of public sentiment and perspective towards GenAI and education.