A Cross-Platform Comparison of Public Discussions Regarding ChatGPT on Twitter and Weibo: Data-Driven Analysis
Zengzhan Xu, Xi Zhang
Abstract
This study provides the first cross-platform comparison of public discussions regarding ChatGPT in two major countries, the United States and China. Representative data from January to July 2024 were collected from Twitter and Weibo, and Python was used for analysis and visualization. Our findings reveal that public discussions about ChatGPT on Twitter and Weibo can be categorized into four consistent themes: technology, products, emotions and education. Among these, education emerged as the most frequently discussed theme. Sentiment analysis shows that both platforms are dominated by positive emotions, with Twitter having a slightly lower proportion (45.2%) compared to Weibo (54.5%). Negative (24.1%) and neutral (30.7%) sentiments on Twitter are marginally higher than those on Weibo (19.9% and 25.6%, respectively). These findings significantly update the previous studies, particularly those suggesting the Chinese public held predominantly negative attitudes toward ChatGPT. Our research also offers valuable insights for policymakers, global technology companies and educators.