Leveraging generative AI in a 3D scenario-based game for vocabulary acquisition and conversation practice: Insights from primary EFL educators through the technology acceptance model

The use of virtual games can provide immersive learning experiences. Advances in generative artificial intelligence (AI) constitute potential of these games for personalised learning. However, many current virtual games have not integrated AI, limiting their benefits. This study addresses this gap by developing a 3D scenario-based game that leverages generative AI for vocabulary acquisition and conversation practice. To evaluate the game, we established a user panel of English as foreign language (EFL) educators, comprising one university English teacher educator and 13 primary school English teachers. To evaluate the generative AI-supported character’s performance in the game, we analysed user-AI dialogue using a multiple linear regression model and visual learning analytics. The participants then evaluated the game through the Technology Acceptance Model. The findings from user-AI dialogue analysis revealed areas for improvement related to the consistency of our generative AI-supported character’s output. Nevertheless, survey results showed the participants’ positive acceptance of the game. The findings from one-to-one evaluations further indicated its potential to support vocabulary acquisition and enhance conversation skills. Finally, the participants recommended additional functionalities, such as audio recording and AI speech generation. This study provides valuable insights into teachers’ needs for generative AI-supported 3D scenario-based games, informing their future development.