earticle

영어

Generative artificial intelligence (AI) is rapidly penetrating the field of language education. This development has opened new avenues for addressing a longstanding challenge in university-level Chinese vocabulary instruction, namely abundant input coupled with insufficient output and weak transferability. Adopting a design-based rather than purely empirical approach, this study focuses on Korean university students majoring in Chinese at the HSK 3–4 proficiency level and proposes an integrated AI-enhanced picture book vocabulary teaching model. The model is grounded in research on AI-assisted vocabulary learning, task-based vocabulary instruction, and multimodal and generative learning theories. Unlike many existing studies that concentrate on early literacy or AI-assisted writing in English, this study targets L2 Chinese vocabulary learning in higher education. Centered on themes such as traditional Chinese festivals, the tasks guide learners to create audio picture books using tools like ChatGPT, AI image generators, and text-to-speech software. The model combines traditional vocabulary explanation with a spiral task cycle comprising five stages: input, application, multimodal output, feedback, and internalization. To minimize over-reliance on AI-generated content and highlight learners’ authentic linguistic efforts, a comprehensive rubric and a “Language Effort Rating Form” are developed and employed. These tools integrate product-oriented and process-focused assessment approaches, thereby making learners’ language investment and revision processes visible. The study then analyzes the model’s feasibility from three dimensions: alignment with teaching objectives, technological accessibility, and teacher–learner readiness. It further provides a detailed lesson plan for the “Dragon Boat Festival (Duanwu Festival)” unit, which serves to illustrate classroom implementation procedures and task workflows. The analysis indicates that the model is highly practical under current university teaching conditions and holds significant potential for broader application. As a conceptual and design-oriented study, it offers a replicable classroom paradigm and design framework for integrating generative AI into vocabulary instruction, while also outlining clear directions for future quasi-experimental studies and classroom-based research.