GenAI-supported multimodal design practices among Chinese primary EFL teachers: an explanatory mixed-methods study of adoption-related factors

Generative artificial intelligence (GenAI) is increasingly influencing English-as-a-foreign-language (EFL) teaching by enabling teachers to generate and adapt instructional resources across multiple modes. However, existing research has mainly focused on AI acceptance and behavioural intention, with less attention to how teachers integrate GenAI into actual multimodal design practices. Drawing on technology adoption perspectives and teacher-related cognitive and affective factors, this mixed-methods study investigated Chinese primary EFL teachers’ GenAI literacy, AI self-efficacy, enjoyment, behavioural intention, and GenAI-supported multimodal design practices. Survey data were collected from 213 in-service primary EFL teachers, followed by semi-structured interviews with 14 teachers. The findings showed that teachers generally demonstrated positive orientations toward GenAI-supported teaching practices. Enjoyment was the strongest predictor of behavioural intention, while AI self-efficacy, behavioural intention, and GenAI literacy were more closely associated with multimodal design practices. This study contributes to GenAI in language education research by extending the focus from technology acceptance to teacher-mediated multimodal design practices and providing insights into how cognitive, affective, and practical factors shape GenAI integration among primary EFL teachers.