Exploring AI as a scaffold for parent–child nature education: a design-based pilot study using SDT-informed semi-structured prompts

Family-based nature education provides important opportunities for children to develop relationships with the natural world, yet many parents face challenges in designing and facilitating meaningful nature experiences in everyday settings. This pilot study explores how generative artificial intelligence (GenAI) can serve as a scaffold for parent–child nature education rather than a substitute for direct engagement with nature or human interaction. Grounded in Self-Determination Theory (SDT), this study developed a semi-structured prompt framework designed to support parents’ needs for autonomy, competence, and relatedness during informal nature exploration activities. A design-based research approach was adopted, involving two iterative design cycles with 11 families across two southwestern Chinese cities. Data were collected through semi-structured interviews and analysis of AI interaction logs. Findings suggest that SDT-informed AI prompts supported parents by providing adaptable activity structures, contextual guidance, and interaction strategies that facilitated parent–child engagement and increased parental confidence. However, challenges were also identified, including AI-generated inaccuracies, task repetition, and differences in parents’ AI literacy. The article also explores strategies for designing GenAI as a relational scaffold that supports parents in facilitating nature experiences while preserving children's direct engagement with the natural environment.