Abstract
This paper investigates the application of Large Language Models (LLMs), specifically DeepSeek, in foreign language education, focusing on the university course “Chinese Culture”. It highlights the advantages of LLMs in enhancing personalized learning, instructional design, and immediate feedback. Through a systematic examination of pre-class preparation, classroom teaching, and post-class extension, the paper demonstrates how LLMs can transform traditional teaching methods by facilitating information retrieval, promoting critical thinking, and offering tailored support to students. The study reveals that the application of LLMs still encounters risks such as technological dependence, which can obscure students’ actual proficiency levels, issues of information credibility, and various ethical challenges. Teachers must address these concerns through curricular restructuring and critical guidance. The research underscores the necessity for foreign language education to strike a balance between technological empowerment and humanistic values. It advocates for teachers to evolve into “AI collaborators” and value guides, enabling them to harness the supportive capabilities of LLMs while preserving the humanistic essence of education.