DOLLama: Fostering Family Anti-Bullying Learning through AI-Augmented, Toy-Mediated Educational Drama

Paper Title

DOLLama: Fostering Family Anti-Bullying Learning through AI-Augmented, Toy-Mediated Educational Drama

Publication Info

• Topic area: AI-enhanced educational drama for anti-bullying learning in family contexts.
• Keywords: AI-powered system, toy-mediated learning, anti-bullying education, family collaboration, educational drama, empathy, gamification, role-play, psychological distance, coping strategies.

Background and Problem

• Problem / challenge: Traditional anti-bullying education relies heavily on teachers and peers in school settings, leaving families without effective tools for engaging children in anti-bullying learning. Parents often lack the expertise to guide such learning, and family settings lack the psychological distance and peer diversity needed for effective role-play.
• Significance: Anti-bullying education is critical for children’s social-emotional development, fostering empathy, confidence, and coping strategies. Extending this education to family contexts could fill gaps in parental knowledge and strengthen parent-child collaboration.
• Motivation and related work: Prior work in educational drama has shown its effectiveness in schools but has rarely explored its application in family settings. Existing HCI approaches, such as VR and video games, lack personalization, dynamic feedback, and mechanisms for involving parents. This paper addresses these gaps by introducing an AI-enhanced system for family anti-bullying education.

Solution

• Proposed approach: DOLLama, an AI-powered, projection-augmented interactive system that uses children’s toys and family-created stories to create gamified anti-bullying vignettes.
• Novelty:
  1. AI-driven co-creation of personalized toy-based characters and stories.
  2. Gamified role-play with AI-controlled characters to foster empathy and practice coping strategies.
  3. Structured phases for learning, role-play, and joint reflection to support family collaboration.
• Procedure and key techniques:
  1. Initiation Phase: Families co-create toy characters and story plots guided by AI, incorporating personal relevance and real-life experiences.
  2. Performance Phase: Children role-play as bullied or bystander characters, interacting with AI-controlled characters in gamified vignettes. Feedback is provided through scoring and narrative adjustments.
  3. Reflection Phase: Families engage in joint debriefing through drawing or letter writing, consolidating learning and exploring real-life applications of strategies.

Results

• Concrete findings:
  • Children learned coping strategies through trial-and-error in gamified role-play.
  • Role-playing as bullied and bystander characters fostered empathy and multi-perspective insights.
  • Reflection activities helped children consolidate strategies and express emotions.
• Advantage over baselines:
  • Enabled personalized, interactive learning at home, unlike traditional school-based or static digital methods.
  • Provided immediate feedback and dynamic narrative adjustments, enhancing engagement and learning outcomes.
• Experiments / evaluation:
  • Conducted a user study with 20 families (43 participants, including 23 children aged 7–11).
  • Data collected through video recordings, interviews, and participant artifacts (e.g., drawings, letters).
  • Analysis revealed patterns of effective family collaboration and insights into parental knowledge gaps.
• Limitations and future work:
  • No quantitative or long-term evaluation of learning outcomes.
  • Limited exploration of cultural and inclusivity considerations.
  • Future work could include longitudinal studies, support for special-needs children, and culturally sensitive adaptations.

Summary

DOLLama is an AI-powered system that transforms children’s toys and family-created stories into gamified anti-bullying vignettes. It enables children to role-play as bullied or bystander characters, co-performing with AI-controlled characters to develop empathy and practice coping strategies. A study with 20 families demonstrated its effectiveness in fostering anti-bullying learning and collaboration, helping parents identify knowledge gaps and discover new educational opportunities. The system’s structured phases and AI-driven features offer a promising approach to extending anti-bullying education into family contexts. Future research could explore broader cultural, inclusivity, and longitudinal dimensions.