A Quest for Information: Enhancing Game-Based Learning with LLM-Driven NPCs

Tereza Tódová

Supervisor(s): RNDr. Vojtěch Brůža

Masaryk University


Abstract: Large Language Models (LLMs) have undergone rapid advancements in recent years. These improvements open new opportunities for creating interactive and immersive learning experiences. We present Quest for Information, a prototype game developed for Virtual Reality (VR) that leverages LLM-based non-player characters (NPCs) to engage players through open-ended, dialogue-driven interactions. By providing contextually appropriate responses, these NPCs help players gather information and complete quests while enhancing immersion and the overall game experience. The primary goal is to demonstrate how LLMs can be used to create dynamic, responsive NPCs that make dialogue-based gameplay more engaging. This work explores various approaches to customizing LLMs to fit the game world, providing tools that can assist developers in building similar educational games. The prototype is designed to be low-cost, easily extendable, and reusable.
Keywords: Video Games, VR
Full text:
Year: 2025