Use of large language models (LLMs) for personalized and accessible mental health support

Abstract

Mental health challenges among students have become increasingly prevalent and concerning in recent years. While traditional approaches to addressing these issues often involve counselling services and therapeutic interventions, there is a growing recognition of the potential benefits of incorporating 3D AI assistants to support students' mental well-being. In the realm of student mental health, where the emphasis is often on stress, anxiety, depression, and academic pressures, the integration of 3D AI assistants and virtual reality technologies offers a novel approach. These AI-driven avatars can serve as personalized virtual companions to students, providing them with emotional support, guidance, and coping strategies. However, the specific software tools and their usability, utility, and acceptance by both students and mental health professionals remain relatively unexplored. To address this gap and harness the potential of 3D AI assistants in the context of student mental health, this paper presents two significant contributions: 1. A comprehensive description of a 3D AI assistant system tailored to the needs of students, offering insights into the technical aspects of avatar creation and customization, thereby laying the groundwork for future research and implementation. 2. An extensive evaluation of the 3D AI assistant system from the perspectives of both developers and students. This evaluation encompasses not only the usability of the technology but also its acceptance and effectiveness as a valuable tool in promoting student well-being. Additionally, this study highlights key findings, shares valuable lessons learned, and acknowledges the primary limitations encountered, paving the way for the potential integration of 3D AI assistants as a transformative resource in enhancing the mental health of students.