간호대학생의 의사소통 훈련을 위한 AI 기반 Virtual Human 시나리오 개발

초록

Background : This study aimed to develop AI-based virtual human scenarios for communication training in nursing education and to evaluate their usability and applicability through expert validation. Methods : This methodological study followed the instructional design model and checklist for AI education. Two scenarios, ‘Admission Management’ and ‘Fall Management,’ were developed. Validation was conducted with 41 experts, including nursing professors and clinical nurses. Quantitative assessments included the Content Validity Index (CVI), Chatbot Usability Scale, and Perceived Virtual Learning Environment Scale. Qualitative content analysis was performed on experts’ open-ended comments. Results : The scenarios demonstrated high content validity, with Scale-level-CVI/Ave values of 0.978 for Scenario 1 and 0.968 for Scenario 2. Usability scores (5-point scale) were generally positive (3.46∼ 4.25), but ‘Time response’ was the lowest (3.46 ± 1.06). Perceived virtual learning environment (7-point scale) showed ‘Confirmation’ rated highest (5.64 ± 1.23), while ‘Immersion’ was rated lowest (4.91 ± 1.46). Qualitative analysis yielded five categories, emphasizing the need for enhanced realism and immersion (e.g., natural voice/avatar), technical refinement (e.g., system stability and response speed), and clearer learning objectives and specific feedback. Conclusion : The developed AI-based virtual human communication scenarios are highly valid and applicable for nursing education. Future development must address technical stability, optimize response time, and improve avatar realism and naturalness to enhance immersion.

키워드