Collaborative hybrid intelligence platform CHIP: A modular architecture for developing and testing personalized lifestyle support interactions

Abstract

The rise of lifestyle-related, non-communicable diseases such as Type II diabetes, cardiovascular diseases, and depression has prompted the development of various behavior change technologies to promote sustained healthy behaviors. User adherence, however, has remained low. The Collaborative Hybrid Intelligence Platform CHIP is introduced to address adherence challenges by placing the user perspective at the center and facilitating dialogue-based interactions between users and their technical and non-technical support systems—including AI systems, clinicians and caretakers. These interactions aim to uncover barriers to adherence and collaboratively shape personalized lifestyle plans that align with a person’s preferences, values, and context. CHIP is a microservice-based research platform written in Python with modules implemented as Docker containers. Its modularity allows researchers to replace or adapt specific components, such as natural language reasoners, for technical evaluation and domain-specific adaptation.