Integrating Generative AI and Immersive Technologies in the Metaverse: Enhancing Learning Experiences through Intelligent Educational Environments
Academic Article in Scopus
This study examines the integration of generative artificial intelligence (AI) tools and immersive technologies within metaverse-based higher education environments, framed by established learning science theories on student engagement [8]-[10]. Specifically, it investigates how the application of large language models (LLMs) such as ChatGPT, MidJourney, and Leonardo AI, combined with Virtual Reality (VR) and Augmented Reality (AR), influences vigor, dedication, and absorption in design learning. The research was conducted within the Tec Virtual Campus, a structured metaverse environment, involving 87 undergraduate architecture and design students in interdisciplinary projects. A mixed-methods approach was employed, combining pre- and post-intervention surveys using the Utrecht Work Engagement Scale (UWES-S-9) and qualitative thematic analysis of student reflections. Statistical results revealed significant improvements (p < 0.01) in all UWES-S-9 subscales, with large effect sizes for dedication (d = 0.88) and medium-to-large effects for vigor (d = 0.76) and absorption (d = 0.73). Qualitative findings showed that GenAI tools supported creative ideation, narrative development, and the overcoming of creative blocks, while immersive validation in VR/AR enhanced spatial iteration and experiential design quality. This paper contributes by (1) offering a detailed methodological account of AI/XR integration in metaverse-based courses, (2) grounding the approach in engagement and experiential learning theories [8]-[10], and (3) providing empirical evidence of its positive impact on creativity and spatial competencies. Limitations include the absence of a control group, the lack of direct performance-based spatial assessments, and the manual, non-structured use of AI prompts. Future work should explore comparative studies with control conditions, advanced prompt engineering, and longitudinal tracking of learning outcomes.
