생성형 인공지능 ‘가상 유아’와의 상호작용 과정을 통한 예비 유아교사의 경험 탐색 = Exploring Pre-Service Early Childhood Teachers’ Experiences Through Interactions with a Generative AI “Virtual Child”

다국어 초록 (Multilingual Abstract)

This study explored the educational potential of generative artificial intelligence(AI) for enhancing pre-service early childhood teachers’ teacher–child interaction competencies through simulated interactions with virtual children. The participants were 30 sophomore students majoring in early childhood education at a university in Korea, and the study was conducted over seven weeks during the first semester of 2025 as part of an early childhood science education course. Using ChatGPT, participants designed virtual children by specifying age, personality, and play contexts, and engaged in interaction practices resembling real teacher–child exchanges. Data were collected from weekly evaluation records, interaction reports, and researcher journals, and were analyzed through iterative reading and categorization. The findings indicated that participants experienced a high level of immersion through unexpectedly natural interactions, engaged in reflective self-assessment of their interaction skills, and developed a willingness to expand the educational use of generative AI. Additionally, participants benefited from opportunities for repeated practice without fear of failure, while also recognizing limitations related to emotional authenticity and interaction continuity. These results suggest that generative AI can serve as an effective educational tool to support the professional development of pre-service early childhood teachers.