Exploring the Role of AI-Assisted Revision with Instructional Scaffolding: A Comparative Study among Korean University Students

• Published in: 언어연구 Vol. 41; no. 1; pp. 41 - 58
• Main Author: Chong Hyein
• Format: Journal Article
• Language: Korean
• Published: 한국현대언어학회 31.05.2025
• Subjects: AI-assisted writing; EFL writing; instructional scaffolding; lexical diversity; process writing; 언어학
• ISSN: 1225-4770; 2671-6151
• DOI: 10.18627/jslg.41.1.202505.41

Exploring the Role of AI-Assisted Revision with Instructional Scaffolding: A Comparative Study among Korean University Students. This study investigated how varying levels of teacher guidance shape the way students revise their writing when using AI support tools. A total of 139 Korean university students engaged in a multi-step writing process, which included planning, drafting, revising with ChatGPT feedback, and final submission. Participants were divided into two groups: one received structured, in-class instruction at each stage, while the other completed the tasks independently with minimal intervention. Quantitative analysis, based on metrics such as word count, sentence length, and lexical variety, revealed that only the supported group demonstrated consistent and statistically significant improvements. In addition, qualitative observations showed that students with classroom guidance engaged more critically and thoughtfully with AI-generated suggestions, often making deeper changes to structure and meaning. These findings suggest that while AI can assist with writing tasks, its benefits are best realized when embedded within a pedagogically scaffolded environment. (Chungnam National University)