Public Attitudes and Spatio-Temporal Characteristics of Robotaxis in China: A Machine Learning-Based Analysis

The pilot operation of robotaxis in China’s public transportation system in 2024 attracted widespread social media attention and posed challenges to the traditional taxi market. During this pilot, public reactions on social media were examined using “Apollo Go” as a keyword; 177,890 relevant social media comments (SMCs) were collected, preprocessed, and annotated. A sentiment prediction model was then fine-tuned and validated for sentiment analysis, followed by topic-based statistical analysis. Results indicate that negative public sentiment toward robotaxis consistently exceeded positive sentiment across all periods. Positive evaluations emphasized regulatory compliance, cost-effectiveness, and technological innovation, whereas negative evaluations centered on driver unemployment, fears of AI displacement, and potential traffic congestion. Sentiment trajectories and the prioritization of concern topics demonstrated temporal and spatial stability. This study provides valuable insights for policy development and market strategies, contributing to the rational planning and efficient operation of autonomous taxis in Chinese cities.