Traditional history education often relies on textbooks and lectures, which limit student engagement with historical events and innovations. The absence of interactive learning methods makes it difficult for students to engage in critical thinking about historical decision-making, technological advancements, and their broader societal impacts.
The Roman Empire Metaverse Simulation, developed in Classlet, introduced an interactive approach to history learning by integrating AI-driven mentorship, scenario-based inquiry, and multimodal learning methods. Students engaged in structured activities, interacting with AI avatars representing Dio Cassius, Seneca, and Emperor Vespasian to explore Roman architecture, ethical trade-offs, and cultural influence. The AI avatars acted as dynamic guides, adapting their responses based on student input and prompting critical thinking.
One student noted, "It's great that they also ask you questions, which leads to more thought and discussion," while another appreciated the game-like interactivity, stating, "It is interesting as the way of interacting with the NPCs is like playing a game." The simulation also integrated 3D object manipulation and artifact analysis, allowing students to examine Roman engineering principles hands-on, rather than passively absorbing content. Through these immersive interactions, students engaged in reflective decision-making, historical comparisons, and experiential problem-solving, strengthening their understanding of Roman society’s complexities.
Students responded positively to the interactive and inquiry-driven format, highlighting its engaging, self-paced nature and stronger retention through gaming-based learning. One participant stated, "It’s easier for me to memorize information through gaming—it’s more immersive and interesting." The integration of AI-driven discussions and multimodal exploration provided an adaptive learning experience, encouraging independent reasoning and personalized engagement with historical content.
Survey results showed ease of use at 4.00/5, engagement at 3.75/5, and perceived usefulness at 3.71/5. However, some students noted challenges with navigation and response lag, suggesting areas for refinement.
By combining AI-driven inquiry, decision-based exploration, and hands-on artifact engagement, the Classlet simulation bridged the gap between passive history education and active historical analysis. Future developments will enhance AI responsiveness, refine interface usability, and expand contextual learning pathways to further support critical thinking, engagement, and adaptability in historical education.
The project is led by Prof. Renia Lopez, with Prof. Phoenix Lam, and Ivan Lau as the co-investigators.
