ABSTRACT As generative AI (GenAI) transforms post‐secondary education, this study explores instructor‐driven approaches to developing AI literacy within mandatory undergraduate language courses in China. Using a convergent mixed‐methods approach, we integrated quantitative survey data from 207 first‐year students with qualitative insights from interviews with 24 students and 6 instructors. Findings indicate that 78% of participants reported an enhanced learning experience, primarily valuing GenAI for immediate feedback and personalized guidance. Predominant uses included comprehension verification (85%) and formative assessment support (60%). Through a utility‐trust gap, students appreciated the accessibility and efficiency of GenAI feedback; however, they expressed reservations about its evaluative accuracy and the potential for over‐reliance. Qualitative analysis further identifies a comfort–agency paradox, whereby increased confidence and reduced anxiety did not automatically translate into strengthened autonomous judgement. Within the ecosystem of AI use, instructor‐driven feedback was valuable for students beyond what GenAI could provide. This study advances an ecological understanding of AI literacy as a negotiated and contextually mediated practice rather than a byproduct of tool exposure. The findings provide empirical guidance for designing instructor‐driven AI literacy initiatives that balance human expertise with generative technologies in assessment‐focused language education.
Delanoy et al. (Sun,) studied this question.