Exploring the impact of nonverbal behavior on educational chatbot user experience

Abstract

3D Digital Humans such as the ones from Unreal Engine Metahumans are becoming used for a variety of tasks from education to health and entertainment. The integration of techniques helps digital humans to engage in dynamic, real-time conversations with realistic human behavior and appearance. In the educational field, digital humans have the potential to significantly enhance student experiences by providing timely information and support. For instance, a well-designed virtual character can efficiently assist students with queries about campus life, admissions, and course content, thereby saving time and improving access to information. Building on this potential, the department of Game and Media Technology at Utrecht University has developed an embodied chatbot aimed at assisting students in the GMT master's program. While this chatbot effectively handles inquiries, its current iteration lacks non-verbal behaviors such as facial expressions, gestures, and gaze—elements that are crucial in natural human communication.

This research aims to explore the impact of incorporating non-verbal behaviors into educational chatbots to enhance user experience. By integrating these behaviors, the project seeks to improve the naturalness and human-likeness of study advisor, making interactions more engaging. The outcomes of this study will provide valuable insights into the design of educational chatbots and contribute to the broader understanding of how non-verbal communication can be leveraged to improve the effectiveness of digital human interactions in educational contexts.