Design and Implementation of a Robot-AI Facilitation System for Supporting Creative Idea Generation

Idea generation critically depends on the facilitator’s personal experience and skills, posing a significant challenge. This study proposes an interactive facilitation support system designed to mitigate such dependency and enable reproducible, productive idea generation for all participants. Conventional human-led facilitation has been criticized for uneven participation, subjective guidance, and the overlooking of diverse perspectives. To address these issues, we designed and developed a robot–AI system that automates core facilitation functions through dynamic question generation employing a humanoid robot integrated with a large language model (LLM). The system dynamically adapts its prompts to guide discussions, ensuring balanced contribution and broad perspective exploration. Preliminary evaluations indicate that our approach can reduce facilitator bias and enhance idea diversity without requiring extensive facilitator expertise. This research demonstrates the feasibility of an automated, dialogue-driven facilitation mechanism that supports reproducible and high-quality ideation processes.