Development of AI-based next-generation psychotherapy agents

Abstract

Although there is a high social need for psychotherapy in the mental health field, the treatment sought by clients is psychotherapy provided by skilled therapists, and the existence of "therapists skilled in psychotherapy" is essential, which limits the number of psychotherapists available. In this context, AI agents have been developed as a technology-based psychotherapy, but their effectiveness does not exceed 60% and is not as high as that of human therapists. This is due to the fact that the purpose of current technology applications is to promote dissemination and implementation, not to enhance efficacy. This study aims to develop and implement an AI agent that is as effective as a human therapist. As a first step, we will develop an AI agent based on the database Retrieval Augmented Generation (RAG), which incorporates the clinical knowledge of therapists skilled in psychotherapy, and prompts. The AI therapist under development will be demonstrated on the day of the event.