초등영어 학습중심평가를 위한 AI 챗봇의 스캐폴딩 유형화 연구

The study on categorizing scaffolding types of AI chatbots for primary English learning-oriented assessment.

The purpose of this study is to explore the types of scaffolding that AI chatbots can provide through systematic feedback based on learners' individual utterances during elementary English learning-oriented assessment. For the study, the researcher designed an English speaking task and implemented an learning-oriented speaking assessment using an AI chatbot for 25 learners in the 6th grade. Subsequently, individual interviews were conducted to analyze the learners' responses to the scaffolding. By analyzing learner responses and transcripts of conversations between learners and chatbots, the researcher categorized the types of scaffolding that AI chatbots can provide during the learning-oriented assessment. As a result, first, the scaffolding can be categorized into three areas: linguistic, cognitive, and affective. Linguistically, the chatbots provided scaffolding for conveying meaning, language forms, and pronunciation aspects. In the cognitive aspect, the scaffolding assisted learners in their thinking processes, while in the affective aspect, it helped regulate frustration. Second, through scaffolding using various strategies, learners demonstrated the ability to self-correct erroneous utterances and gradually complete sentences on their own. In addition, cognitive scaffolding helped learners solve challenging tasks at their current cognitive level, and affective scaffolding contributed to improving confidence in speaking English. The study offers implications and recommendations based on the findings.