Application Risks of Artificial Intelligence Chatbots in Psychotherapy and Feasible Solutions for Legal and Policy Issues

Abstract

This paper reviews existing literature on therapeutic misconception in AI mental health chatbots, a fast-growing field that integrates digital technology with mental health services. As AI-driven psychological intervention tools become increasingly widespread in clinical and daily settings, the applicability of traditional ethical frameworks to human-AI psychological interaction has sparked intense academic debate. This review synthesizes relevant empirical and theoretical studies, compares competing ethical perspectives, most notably the traditional Therapeutic Misconception (TM) theory and the alternative framework, and then systematically examines the consistencies, contradictions, emerging trends, and inherent limitations in current research. It aims to clarify the theoretical disputes surrounding therapeutic misconception in AI mental health contexts, identify key gaps in the existing body of knowledge, and lay a solid foundation for formulating rational norms that balance technological innovation and user rights protection. Furthermore, the paper puts forward targeted legal and policy outlooks, drawing on typical global regulatory practices, and proposes feasible solutions to address current regulatory ambiguities, responsibility vacuums, and inadequate protection for vulnerable groups, thereby providing practical reference for the healthy development of the AI mental health industry and the effective safeguarding of user interests.