본문 바로가기
  • Home

LLMs for Living Chinese Teaching Materials: A Pedagogical Transformation of Social Media Texts

  • Journal of Chinese Language and Literature
  • 2026, (101), pp.439~463
  • DOI : 10.15792/clsyn..101.202604.439
  • Publisher : Chinese Literary Society Of Yeong Nam
  • Research Area : Humanities > Chinese Language and Literature
  • Received : March 20, 2026
  • Accepted : April 13, 2026
  • Published : April 30, 2026

Minjun PARK 1

1덕성여자대학교

Accredited

ABSTRACT

This study explores a methodology for converting Chinese social media texts into educational materials using Large Language Models (LLMs) and proposes an interactive platform called ChineseRealTalk. In Chinese language education, there has been a persistent gap between the language presented in textbooks and the authentic everyday language used by native speakers on social media platforms. Despite the pedagogical value of social media texts that reflect the linguistic sensibilities of contemporary learners, incorporating such materials into the classroom remains challenging, as manually adapting them to match learners’ proficiency levels and preparing customized learning texts require considerable time and effort. To address this issue, this study developed ChineseRealTalk by leveraging the contextual understanding and text generation capabilities of LLMs. The platform offers three core functions: Level-based Text Converter, which rewrites original texts according to HSK levels; Neologism and Difficult Expression Annotator, which provides pinyin and example sentences for unfamiliar expressions; and Comprehension Check Question Generator, which automatically produces multiple-choice quizzes and discussion topics. While the platform significantly reduces teachers' preparation time, the study also identifies several limitations, including LLM hallucinations, the loss of cultural context, and delays in reflecting the latest linguistic trends due to knowledge cutoff. Ultimately, LLMs function most effectively as draft generators for instructional materials. However, their effective use depends on teachers’ active curatorial competence, including the critical evaluation of generated outputs and the careful design of prompts.

Citation status

* References for papers published after 2024 are currently being built.