Integrating ChatGPT-powered AI tutoring with virtual reality classrooms creates highly engaging and well-tolerated learning environments for cardiac anatomy education in K-12 students.
Objective: The main goal of this study was to design, develop, and evaluate a social virtual reality (VR) classroom system that integrates text-based conversational AI (ChatGPT) to enhance cardiac anatomy education for K-12 students. The researchers aimed to investigate how the combination of immersive 3D visualization, collaborative learning environments, and intelligent AI tutoring could address challenges in teaching complex anatomical concepts to younger learners while making high-quality education more accessible and affordable for schools with limited resources.
Methods: The study employed a comprehensive mixed-methods approach involving system development and user evaluation. The researchers developed a social VR classroom using Unity and Photon platforms, featuring a detailed interactive 3D heart model that multiple students could explore simultaneously. The system integrated ChatGPT through OpenAI's API to provide real-time, text-based conversational support. Fifteen K-12 students (ages 11-24, with diverse educational backgrounds from primary school to bachelor's degree level) participated in 30-40 minute VR sessions using Meta Quest 2 headsets. The evaluation utilized a multi-dimensional framework consisting of four standardized questionnaires: Learning Perception Questionnaire (LPQ), VR Perception Questionnaire (VRPQ), AI Perception Questionnaire (AIPQ), and VR Symptoms Questionnaire (VRSQ). Each instrument used 5-point Likert scales to assess different aspects of the learning experience, from perceived effectiveness to potential adverse effects.
Key Findings: The results demonstrated overwhelmingly positive outcomes across all measured dimensions. Learning perception scores were significantly above neutral (M = 3.96), with 100% of participants agreeing that interactive tools were effective for learning heart anatomy. VR perception showed exceptionally high ratings (M = 4.53), with all participants rating VR integration as useful for cardiac anatomy learning and 86.7% agreeing it was more effective than traditional 2D materials. AI perception similarly demonstrated strong positive responses (M = 4.51), with 100% of participants finding AI explanations clear and understandable. Importantly, VR-related symptoms were significantly below the neutral point (M = 2.00), indicating minimal adverse effects. Specifically, 86.7% experienced minimal dizziness, 73.3% reported minimal eye strain, and 66.7% experienced minimal disorientation. All questionnaire scales demonstrated acceptable to excellent internal consistency (α = .71-.85), supporting the reliability of the measurement instruments.
Implications: The findings contribute significantly to the field of AI in education by demonstrating the synergistic potential of combining multiple emerging technologies. The research provides compelling evidence that integrating conversational AI with immersive VR can create highly engaging learning environments that address multiple learning modalities simultaneously. The unanimous positive ratings suggest that this approach could revolutionize STEM education by making complex subjects more accessible and engaging. The study supports constructivist learning theories by emphasizing active engagement in knowledge construction and demonstrates that modern AI can provide personalized, adaptive explanations that accommodate diverse educational backgrounds. This has particular implications for K-12 education, where engagement and accessibility are primary concerns, potentially providing a pathway to more inclusive and effective science education across various educational contexts.
Limitations: Several important limitations constrain the generalizability of these findings. The small sample size (N = 15) limits statistical power and broader applicability. The absence of a control group prevents direct comparison with conventional instruction methods, making it difficult to isolate the specific benefits of the VR-AI integration. The study measured only perceptions and user satisfaction rather than actual learning gains or knowledge retention, which represents a significant gap in understanding educational effectiveness. The single-session design doesn't address long-term learning outcomes or sustained engagement. Additionally, the diverse age range and educational backgrounds of participants may have introduced confounding variables that weren't adequately controlled.
Future Directions: The researchers recommend several critical areas for future investigation. Randomized controlled trials comparing the VR-AI system to conventional anatomy instruction are essential to establish true educational effectiveness. Studies should incorporate pre-post knowledge assessments and longer-term retention measurements to evaluate actual learning outcomes. Research on optimal session duration to balance learning gains against potential fatigue is needed. Investigation of adaptive AI tutoring strategies that adjust to individual learning progression would enhance personalization. The incorporation of biometric feedback could provide objective measures of engagement and cognitive load. Future studies should also explore the scalability of such systems across different educational contexts and anatomical topics beyond cardiac anatomy.
Title and Authors: "Exploring 3D Cardiac Anatomy with Text-Based AI Guidance in Virtual Reality" by Fatima-Ezzahra Boubakri, Mohammed Kadri, Fatima Zahra Kaghat, Ahmed Azough, and Hamid Tairi.
Published On: 2025
Published By: 2025 International Conference on Smart Learning Courses (SCME), IEEE