Immersive Learning in History Education: Exploring the Capabilities of Virtual Avatars and Large Language Models

Abstract

This paper introduces a novel approach to K-12 history education through the development, implementation, and evaluation of ”Artificial Legacy in Virtual Environments” (ALiVE), an immersive chatbot powered by a Large Language Model (LLM). ALiVE simulates interactive conversations with historical figures such as Albert Einstein, Marie Curie, Napoleon, and other lesser-known individuals, utilizing various historical sources, including speeches, letters, and diary entries. This innovation addresses the pressing need for engaging educational tools that effectively incorporate historical accuracy and contextual information. ALiVE enhances learning by enabling students to engage in fictional dialogues with these figures, represented as animated NPCs. The system also includes a text-to-voice function that mimics the real-life voices of historical figures, further enriching the user experience. Our evaluation comprises two studies: a system evaluation that assesses the correctness and relevance of the system’s responses using various metrics and
an educational evaluation in which history teachers assess the system’s affordances, challenges, and potential for additional functionality. The combined results from these evaluations demonstrate the educational value and feasibility of using immersive virtual environments and LLM based chatbots in history education, highlighting both the opportunities and challenges. We also present the first heuristic model for Immersive History Education with LLMs (iHELLM), which integrates the role of historical sources and contexts in enhancing both physical and virtual learning environments. Future directions focus on improving educational scenarios for deeper curricular integration and developing more immersive interactions with virtual avatars to significantly enhance both the quality of and engagement in historical education.