<p>This study aimed to investigate the factors accounting for Chinese teachers’ concerns about teaching artificial intelligence (AI). Based on the concerns-based adoption model and the pedagogical content knowledge framework, a hypothesized model associating teachers’ knowledge, perceived social good, and concerns about teaching AI was tested via structural equation modelling. The responses from 269K-12 AI teachers in southern China were utilized to test the hypothesized model. Structural equation modelling reveals that the association between teachers’ knowledge of teaching AI and teachers’ concerns about teaching AI is mediated by teachers’ perceived social good of teaching AI. Particularly, teachers’ perceived social good of teaching AI partially mediated relationships between teachers’ pedagogical AI knowledge and refocusing concern, as well as teachers’ conceptual AI knowledge and management concerns. These findings provide a more profound understanding of teachers’ perceived social good as a pedagogical belief. The results show that teachers’ knowledge (i.e. pedagogical AI knowledge and conceptual AI knowledge) predicted higher stages of concern (i.e. refocusing and management) when mediated by teachers’ perceived social good of teaching AI. This study contributes to a better understanding of factors contributing to teachers’ concerns about teaching AI, and how to address them for teacher professional development.</p>