국내 대학생 대상 AI 기반 평가도구 연구 동향 분석: 주제범위 문헌고찰

초록

Purpose This study aimed to systematically analyze domestic research on AI-based assessment tools for university students using a scoping review. Methods A scoping review was conducted using a five-stage framework. Studies published from 2020 to August 2025 were searched in RISS, KISS, KCI, DBpia, CINAHL, PubMed, Web of Science, EBSCO, and ERIC. Among 194 records, 13 studies met the inclusion criteria. Results The studies were published between 2022 and 2025 and were synthesized into three themes: functional validity and limitations, qualitative characteristics and educational utility of AI feedback, and sociotechnical acceptance and ethical issues. AI-based tools showed some agreement with human evaluation in areas with relatively clear criteria, such as grammar, vocabulary, and sentence structure. However, limitations were found in assessing creativity, critical thinking, and logical development. Ethical issues included source attribution, academic integrity, dependence on AI feedback, and instructor review. Conclusion AI-based assessment tools may complement university assessment, particularly in structured tasks and immediate feedback. Further validation is needed across disciplines, assessment tasks, and Korean-language educational settings, with clear rubrics, AI prompts (instructions provided to AI systems), and ethical guidance.

키워드