Development of an LLM-based Conversational Blockchain Serious Education Game

다국어 초록 (Multilingual Abstract)

Blockchain technology has been applied across various domains including finance, logistics, and healthcare. However, its core concepts such as distributed ledgers, consensus mechanisms, and smart contracts remain abstract, hindering comprehension among non-experts. Particularly in everyday services like local currency gift certificates, citizens perceive them merely as payment methods, making it difficult to grasp structural meanings such as immutability, transparency, and conditional execution. Therefore, this study developed a Unity 2D mobile serious game framework that integrates an LLM-based natural language conversational interface with a lightweight blockchain simulation engine. A State Manager was applied to mitigate the interaction constraints arising from the stateless nature of LLMs, and local currency transaction scenarios were employed to support learning blockchain concepts within real-life contexts. The development achieved an average LLM response time of 2.3 seconds, a command recognition rate of 93.3%, and stable performance above 35fps in mobile environments, thereby presenting a method to lower the entry barriers to blockchain learning through natural language-based interaction and practical scenarios.