Experts' Perception of AI-Generated Personalised Stories for Health Promotion.

A qualitative study is presented aimed to explore how person-tailored stories generated by a storytelling system agent built on a combination of argumentation theory and large language models (LLM) can support a person in reflecting on activities promoting health and potential lifestyle changes. We conducted a formative evaluation with ten domain experts to explore perceived purposes, benefits and risks. They saw both potential opportunities, if applied with a clear purpose for their target groups, and risks in using LLM-based stories for health reflection if not carefully designed to adapt to an individual's needs. The results will guide further design and implementation of the storytelling functionality, and studies involving older adults.