ABSTRACT Although generative artificial intelligence (GenAI) literacy models have recently emerged in the field of second language (L2) teaching, most of them adopt a top‐down approach by modifying existing frameworks, such as digital literacy models. As a result, these models may not fully capture the contextually grounded competencies needed for effective L2 teaching. Furthermore, many of these models are designed for in‐service L2 teachers, leaving pre‐service teachers largely unaddressed. To fill this gap, this study utilised the Behavioural Event Interview (BEI) to analyse the GenAI literacy of 20 pre‐service L2 teachers from universities in China, who were divided into two groups: 10 high‐performing and 10 average‐performing participants. Using MAXQDA 2022 for data analysis, 18 key GenAI literacy attributes were identified. An independent t ‐test conducted with SPSS 26 revealed that nine of these attributes were discriminative (i.e., Agency , Developmental Motivation , Resilience , Prompt Design , Multiple‐tool Integration , Classroom Interaction , Personalised Teaching , Reflective Practice , and Collaborative Innovation ), distinguishing high performers from average ones. The remaining attributes were categorised as benchmark competencies (i.e., Value Cognition , Characteristic Cognition , Basic Operations , Resource Preparation , Interdisciplinary Goal Integration , Content Evaluation , Homework Grading , Ethics , and Emotional Care ), which are essential for all pre‐service L2 teachers. Based on McClelland's Iceberg Model, these attributes were further organised into four dimensions: Knowledge , Skills , Traits , and Motives . Finally, we discuss the implications of our model and provides directions for future research in GenAI literacy for pre‐service L2 teachers.
Wu et al. (Tue,) studied this question.