Designing for Context: The Interplay of AI Agent Types and Situational Urgency in VR Autonomous Vehicles

This study investigated how AI agent type and situational urgency jointly influence user experience in virtual reality autonomous vehicles (AVs). Using a 3 × 2 mixed factorial design, participants interacted with three agent types (voice-only, robot-like, and human-like) across two levels of urgency (non-emergency and emergency). Seven psychological variables, including presence, workload, perceived benefit, perceived risk, interpersonal feelings, behavioral intention, and attitude, were evaluated. Results revealed significant interaction effects for perceived benefit, interpersonal feelings, and behavioral intention. Notably, the robot-like agent in emergency scenarios elicited the highest perceived benefit and behavioral intention, while the voice-only agent produced the most positive interpersonal feelings in the same context. These findings suggest that contextual appropriateness, rather than the level of anthropomorphic realism, plays a more critical role in shaping user engagement and behavioral responses in AV contexts.