Beyond the Basics:Analyzing the Evolution of Prompt Literacy in KFL Learners’ ChatGPT-Assisted Writing

This study investigates the characteristics of prompt literacy observed in intermediate-level Korean language learners at a university in Australia during writing tasks assisted by AI Chatbot (ChatGPT). Over the course of one semester, students completed three writing assignments, using ChatGPT to revise and refine their texts. During this process, interaction logs, consisting of learner generated prompts and corresponding AI responses, were collected for analysis. Qualitative analysis revealed three key characteristics in learners’ prompt usage: uniformity (repetitive use of limited prompt structures), discreteness (lack of continuity in interactions), and passivity (uncritical acceptance of AI responses). These traits indicate that learners tended to use ChatGPT primarily as a translation or editing tool, rather than as a dynamic conversational partner, thereby underutilizing its educational potential. To gain a more nuanced understanding of learners’ linguistic patterns and cognitive approaches in interacting with AI, a text mining analysis was conducted on the collected prompts. This quantitative approach enabled the identification of frequently used expressions, types of questions, and structural tendencies, offering deeper insights into learners’ language habits, patterns of thinking, and strategic engagement with ChatGPT. By combining qualitative and quantitative methods, this study provides empirical evidence for the need to support prompt literacy development in language education. The findings highlight the importance of instructional strategies that foster more interactive, critical, and effective use of generative AI tools in Korean language learning contexts.