생성형 AI 활용 대학교육에 대한 교수자의 인식: 활동체제 모형을 중심으로

초록

As generative AI rapidly spreads in higher education, instructors express both expectations and concerns about AI integration. This study analyzed expected benefits and challenges of generative AI-integrated higher education systems using activity system theory. Semi-structured interviews were conducted with 10 instructors from University in Seoul who had experience using generative AI, and data were analyzed using thematic analysis. Instructors expected generative AI to enhance students’ meta-cognition, provide personalized education, and support cognitive expansion. However, challenges from contradictions between activity system components were identified: cognitive offloading (subject-tool contradiction), knowledge distortion from AI hallucinations (tool-object conflict), undermining of critical thinking and academic literacy culture (community-object contradiction), and constraints on instructors’ feedback and assessment authority (object-division of labor contradiction). To address these challenges, instructors proposed learner-centered AI-based learning design and strengthening human services. This study confirms that generative AI integration requires structural changes across the entire higher education system beyond technical implementation. Based on the findings, this study provided practical implications of establishing instructor-learner-AI triangular collaboration systems, providing instructor-AI hybrid feedback, utilizing AI query logs for learning analytics, fostering students’ critical AI literacy, and developing discipline-specific AI-integrated instructional models. Effective integration of generative AI in universities requires structural improvements from a systemic perspective beyond individual capacity building.

키워드