An Architecture for Interactive Storytelling in the Domain of Care Training

This chapter delves into the transformative potential of AI-driven narrative systems in healthcare training, particularly for managing eating disorders like anorexia nervosa. It introduces a hybrid AI engine that integrates story graphs, rule-based logic, and Large Language Models (LLMs) to create dynamic and adaptable narrative frameworks. The chapter explores the challenges of traditional serious games and the limitations of conventional game engines, emphasizing the need for flexible, personalized, and contextually rich learning experiences. It presents a use case involving family-based therapy (FBT) for anorexia nervosa, demonstrating how interactive digital storytelling can enhance caregiving competencies and emotional resilience. The chapter also outlines the requirements for designing effective interactive storytelling systems, including experiential learning, engaging storytelling, authorability, and customization. It discusses the limitations of existing systems and the potential of LLMs to generate coherent, context-sensitive responses in real time. The chapter concludes with a detailed example of an interaction scenario, illustrating how the hybrid AI engine can be used to train caregivers in supporting adolescents with eating disorders. The chapter highlights the importance of continuous integration and reusability, making it a valuable resource for professionals interested in the intersection of AI and healthcare training.