Beyond Reality: Enhancing Engagement and Contextual Learning with Generative AI Agents in the Metaverse Across UGC Universities

In this presentation, led by the Principal Investigator from Lingnan University’s Science Unit, we explore scenarios from the "Beyond Reality" FITE project, showcasing contributions from co-PIs across UGC universities. This project integrates Generative AI agents within Classlet, a metaverse platform, to develop immersive, interactive learning environments across fields such as linguistics, computer science, interactive media, and food science. Scenarios demonstrate AI agents acting as realistic guides in various subjects. At CUHK, linguistics students engage with AI agents to develop research method skills; at HKUST, agents help computer science students explore audio compression; at HKBU, students work with agents to optimize search engine algorithms; and at PolyU, food science students interact with agents in virtual food production scenarios, learning about sustainable resource management. These agents adapt to student input, promoting critical thinking and applied learning. While findings show that AI agents enhance engagement, further development is needed to create deeper context, situational awareness, and defined roles within scenarios. Key engagement constructs like Perceived Usefulness and Intent to Use still account for significant engagement variance (e.g., 64.2% in selected modules), but the addition of nuanced contextual cues and role-based interactions is essential to drive more effective learning outcomes. The "Beyond Reality" project’s expansion across UGC universities illustrates the potential of AI-driven, context-rich environments for real-world skills development.