Beatrice: A Chatbot for Collecting Psychoecological Data and Providing QA Capabilities

Abstract

In recent years, the continuous advancement of artificial intelligence technology has generated increasing interest in interactive psychological dialogue systems as a means of providing psychological health assistance based on AI technology. This paper introduces Beatrice, a chatbot specifically designed to address the challenge of obtaining psychological ecological data in intensive longitudinal designs (ILDs). Beatrice is a low-code interactive dialogue agent that can be seamlessly integrated into various software platforms and abstracts psychological scales. By leveraging large pre-trained language models and prompt learning, we have achieved a direct and effective psychological support system that enables natural interaction with a diverse range of participants, facilitating the acquisition of psychoecological data. Additionally, we conducted a small-scale system usability test and interviews based on the Technology Acceptance Model (TAM) framework, thereby providing an opportunity to investigate the emotional regulation mechanisms in Chinese adolescents. Through iterative processes and optimization, this tool aims to enhance the application of ILDs by providing improved support.