Age Matters: Age-Differentiated Playtesting for AI Educational Games

Playtesting is widely employed in educational game design as a post-hoc method for game design validation within the age groups that the learning goals target. For domains such as AI literacy, age-targeted educational games testing may be limiting, as AI learning goals are relevant and frequently misunderstood by children, adolescents and young adults. Designing educational games that teach AI to a wider audience of ages requires playtesting with players of different ages to better interpret gameplay and narrative appropriateness of AI in AI educational games. In this study, we playtested four AI educational games with middle, high school, and college aged students, and analyzed playtesting data, including verbal discourse, in-game behaviors, and post-playtesting reflections. Our findings demonstrate that playtesting can serve as an asset-based practice which reveals how players of different age groups make sense of educational games during play. We provide critical reflections from multi-age playtesting of AI educational games that inform design, and recommendations for engaging AI educational game design.