AI-Enhanced Virtual Reality University Office Environments: An Empirical Study of Student Perceptions, Presence, and Adoption Intent

Abstract

Integrating artificial intelligence and virtual reality technologies will influence how we teach and learn at the university. Furthermore, very few studies examine students’ views of AI-driven virtual reality simulations of university office settings. We present a study at International Balkan University with 58 pre-session participants and 29 post-session observers, involved in a virtual reality simulation of the university office scenario using an AI-integrated assistant. This study evaluates five constructs (observation quality, presence and realism, artificial intelligence feature perception, usability and adoption intent, and future readiness) using validated Likert-scale questionnaires and open-ended qualitative responses. Our findings indicate high accuracy in perceiving artificial intelligence (M = 4.48/5), remarkably high clarity of observation (M = 4.39/5), and moderately high adoption intent (M = 4.12/5). In contrast, privacy concerns (M = 3.24/5) and trust calibration are still problematic areas. Pre-session findings reveal that most participants had no meaningful virtual reality exposure (M = 1.88/5); however, they were quite experienced with artificial intelligence tools (M = 4.26/5), defining a sharp technology asymmetry in the sample. Qualitative comments show that artificial intelligence meeting summarization and instant question/answering are the most favored features, while data privacy issues and academic over-reliance are the main concerns. This study introduces a new survey and mixed-methods approach to the growing scholarly discussion of the metaverse as an educational platform. At the same time, it reveals direct implications for the design of AI-powered virtual reality spaces for learning.