Will ChatGPT reshape society? An analysis of multiple social media platforms-based ChatGPT comment themes, sentiment evolution, and development trends

This article provides new directions for understanding the changes in AIGC (Artificial Intelligence Generated Content) technology and users’ perceived needs. Meanwhile, this article provides theoretical and practical references for the application, development and governance of ChatGPT by related enterprises, organisations and users. In this article, we first use Srcapy crawler technology to crawl 31,377 text data from Sina Weibo, X (Twitter), and Tumblr comments on the keyword ‘ChatGPT’. Second, this article analyses the theme evolution of multiple social media platforms’ comments on ChatGPT themes based on the life cycle theory and the latent Dirichlet allocation (LDA) theme model. Third, this article uses sentiment analysis to explore further the process of sentiment classification and sentiment evolution analysis of ChatGPT comments. Finally, this article predicts the future development trends of ChatGPT in terms of social influence through the analysis of theme sentiment crossover and evolution. It is found that the trend of comment theme evolution about ChatGPT on multiple social media platforms is in line with the life cycle theory, which can be categorised into four phases: latent period, hot discussion period, stable period and decline period. It is known that the public sentiment generated towards ChatGPT can be categorised as positive, negative and neutral. ChatGPT is greatly influencing several aspects of society, especially in the field of education, and will have better development and application prospects in many social industries. ChatGPT-like technologies will promote the deep transformation of the social structure and contribute to the prosperity of the digital economy and the diversification of the cultural industries. Finally, this article puts forward suggestions for the development of ChatGPT.