A Parent-Child Storytelling Model with Generative AI for Emotion Learning

Generative AI (GenAI) is increasingly integrated with storytelling to cultivate children’s emotion competence. However, most GenAI storytelling practices remain limited to passive content delivery, lacking scaffolded questions that encourage emotional reflection, and the role of parents in collaborating with GenAI is overlooked. This study examined how different GenAI-supported storytelling models influence first-graders’ emotion competence in the context of separation anxiety. Seventy-five first-graders were randomly assigned to three groups: Standard Storytelling GenAI (n = 25), Dialogic Storytelling GenAI (n = 25), and Parent-Assisted Dialogic Storytelling GenAI (n = 25). Data were collected through an emotion competence observation scale and records of child–GenAI dialogues and interactions. Results indicated that the Parent-Assisted Dialogic group significantly outperformed the other two groups in emotion perception and identification, understanding, and regulation. These findings underscore the critical role of parental scaffolding, categorized into four types: emotional support, cognitive scaffolding, translation and clarification, and emotion labeling.