본문 바로가기
  • Home

Prompt Design for AI-Assisted Writing Instruction for the New HSK Level 6 ― An Exploratory Study

  • The Journal of Study on Language and Culture of Korea and China
  • Abbr : JSLCKC
  • 2025, (77), pp.3~24
  • DOI : 10.16874/jslckc.2025..77.001
  • Publisher : Korean Society of Study on Chinese Languge and Culture
  • Research Area : Humanities > Chinese Language and Literature
  • Received : July 10, 2025
  • Accepted : August 20, 2025
  • Published : August 31, 2025

Kim Seon-Ah 1

1용인대학교

Accredited

ABSTRACT

This study recognized the problems of traditional teaching methods in Chinese writing education and proposed the necessity of AI-powered personalized learning. In particular, it suggested AI prompt-based teaching and learning methods to overcome the limitations of New HSK Level 6 writing education and improve learning efficiency. The research results are summarized as follows: First, New HSK Level 6 writing was confirmed to be a complex task that comprehensively evaluates reading comprehension, memory, summarization, and writing skills, requiring students to read and memorize a 1,000-character passage for 10 minutes, then summarize it in approximately 400 characters within 35 minutes. Second, it was found that New HSK Level 6 writing question types can be categorized into event-centered, character-centered, and fable/classical allusion types, with each type requiring differentiated approach strategies. Third, this study proposed AI-integrated teaching methods by designing specific prompts for reading, writing, and evaluation domains. In the reading domain, prompts were presented for understanding the overall plot, key elements, and narrative structure. In the writing domain, detailed prompts were designed for grammar correction, sentence improvement, and paragraph optimization. In the evaluation domain, prompts reflecting comprehensive criteria including content relevance, logical coherence, grammatical accuracy, and format compliance were proposed.

Citation status

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