Author Information
Cheng-Yuan Wei, National Taiwan University of Science and Technology, TaiwanYu-Chi Chen, National Taiwan University of Science and Technology, Taiwan
Chih-Chung Chien, National Taiwan University of Science and Technology, Taiwan
Huei-Tse Hou, National Taiwan University of Science and Technology, Taiwan
Abstract
In recent years, the application of generative AI's (GAI) adaptive characteristics in education has rapidly emerged. However, general GAI in educational games often lacks realistic situations and operational fidelity, leading to limited authentic experiences and difficulties in learning transfer. Additionally, GAI frequently produces inaccurate or off-topic responses. To address these limitations, this study designs a historical education problem-solving game based on a previous research team's framework of using GPT as a game Non-Player Character (NPC) (Chen & Hou, 2024). The game incorporates realistic story situations and reduces off-topic NPC responses. In the game, learners play as the close friend of the protagonist and collaborate with GPT-based NPC peers designed by the research team. They conduct online information searches within a limited time to find the historical period in which the game protagonist disappeared and the name of the current building to save the protagonist. A total of 17 participants engaged in the empirical evaluation of this study. The study found that learners exhibited a high flow state, perceived the game as highly playable and enjoyable, and had a high acceptance of the game. Additionally, 70% of learners believed that the game helped with historical learning, and nearly 50% felt that it was closer to real human interaction compared to typical GPT conversations. The study demonstrates that a GPT-based NPC, enhanced with contextual stories and reduced off-topic responses, can effectively improve learners' gaming experience.
Paper Information
Conference: ACE2024Stream: Design
This paper is part of the ACE2024 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window
To cite this article:
Wei C., Chen Y., Chien C., & Hou H. (2025) Combining Realistic Story Situations and GPT-Based NPC Framework for Historical Knowledge Problem-Solving Games ISSN: 2186-5892 – The Asian Conference on Education 2024: Official Conference Proceedings (pp. 257-262) https://doi.org/10.22492/issn.2186-5892.2025.23
To link to this article: https://doi.org/10.22492/issn.2186-5892.2025.23
Comments
Powered by WP LinkPress