Chatbots in Mental Healthcare: Developments and Challenges

In this paper, we employ systematic bibliometric analysis to reveal the progressive increase in chatbot research within mental healthcare. Breakthroughs in artificial intelligence have revolutionized chatbot interaction capabilities and catalyzed a paradigm shift in mental healthcare from traditional unidirectional diagnosis and treatment to human-machine collaborative models. Synthesizing existing literature with the evolution of natural language processing from rule-based to pre-trained language models, this paper categorizes the development of mental healthcare chatbots into four distinct phases: preliminary exploration, machine learning era, deep learning era, and large language models era. Based on current trends, this study identifies critical challenges and aims to provide insights for advancing chatbot technology, constructing ethical frameworks, and optimizing human-machine collaboration in mental healthcare. Future work may focus on paradigm shifts in research, such as incorporating embodied intelligence to advance the “humanization” process of mental health service chatbots, and the establishment of robust regulatory mechanisms to ensure their safety and effectiveness.