Advancing Patient-AI Conversations with Recursive Learning Memory Using Relational Frame Theory

The future of Artificial General Intelligence (AGI) in bedside care relies on integrating human psychological principles to foster advanced cognitive abilities such as reasoning, problem-solving, and spatial awareness. Key features like recursive learning memory and human-like context awareness allow AI systems to continuously learn and recall patient information in a relational manner. To achieve this, we propose an AI framework based on Relational Frame Theory (RFT), which organizes patient data to reflect human relational patterns. This approach enables iterative and context-sensitive information retrieval through a dynamic knowledge graph. Our evaluations demonstrate that this method enhances patient interactions, offering deeper, more personalized engagement that surpasses traditional Retrieval Augmented Generation in its ability to emulate nuanced, human-like understanding.