Narrative Co-Construction with Conversational AI in Augmented Reality: A Conceptual Framework for Collaborative Story-Making

Abstract

Augmented reality (AR) and conversational artificial intelligence (ConvAI) have been explored as design components in interactive learning environments for children, particularly in literacy and storytelling contexts. However, most existing systems position artificial intelligence (AI) as a narrator, with learners primarily acting as recipients of pre-authored content. This interaction model limits learner agency and constrains opportunities for shared meaning-making. To address this gap, this paper introduces narrative co-construction, in which children and AI collaboratively create and shape stories through ongoing interaction. Narrative co-construction is conceptualised as a shift from storytelling to story-making, emphasising shared control, turn-taking, and dynamic narrative development within AR learning environments. Building on interaction design and child-centred learning principles, we propose a conceptual framework, the Collaborative Narrative Interaction Framework (CNIF), which outlines key components of co-constructive narrative interaction, including narrative initiative, conversational responsiveness, and adaptive story branching. The framework is illustrated through AR-based learning scenarios that demonstrate how ConvAI can function as a collaborative partner rather than a narrative authority. By reframing the role of AI from content provider to co-author, this work contributes design-oriented insights for developing AR learning environments that support creativity, engagement, and active meaning-making in children’s learning experiences.