Exploring Public Perceptions of Generative AI in Education with BERTopic-based Topic Modelling and LLM-driven Sentiment Analysis

Abstract

Generative AI, since 2022 when ChatGPT was released, has been widely used in education. While bringing tremendous benefits to teachers and students, it also poses numerous challenges. Exploring the public's opinions and sentiment towards Generative AI in education will greatly promote their integration. Weibo, as the most important social media in China, offers a wealth of data to analyze public attitudes and viewpoints. In this study, 17,799 weibo posts was crawled for topic modelling and sentiment analysis by using BERTopic algorithm and the Large Language Model DeepSeek respectively. It's suggested that the public mainly focus on seven topics related with Generative AI in education, including AI-driven teaching innovation and digital transformation, applications of AI in academic research and the resulting changes in academic publishing, applications and experiences of AI in language education, risks, challenges of using GenAI in education and responses, feedback and complaints after using AI, the development of artificial intelligence major and related fields as well as the impact of AI on the job market and responses. Nearly half users embrace this new technology for its convenience, while some people demonstrate different attitudes. They worry such risks as the job loss, plagiarism, over-reliance and so on. At the end this study, several suggestions are put forward to tackle those negative sentiments to promote the the deep integration of Generative AI in education. Meanwhile, the limitation of this study was also proposed.

Ähnliche Arbeiten