Perceived benefits and limitations of a generative AI chatbot for mental health support: an exploratory mixed-methods study

This study explores the perceived benefits, limitations, use patterns, and future relevance of the “Psychologist” chatbot on Character.AI, utilising a mixed-methods approach involving online semi-structured interviews and surveys. Amidst a growing need for accessible mental health resources, this research compares a generative AI-driven chatbot’s unique position, with traditional, non-generative AI mental health tools. Quantitative data from 13 survey responses indicated a significant young female demographic (primarily 18-24 years) engaging with the chatbot, revealing patterns of weekly usage and varying session lengths. Qualitative analysis of eight interviews unveiled three primary themes: the perceived therapeutic attributes of the AI’s persona, the empowerment of user agency, and enhanced accessibility. Key factors in the chatbot’s appeal included its perceived ability to provide empathy, non-judgment, and validation, alongside its 24/7 availability and cost-free access. This study underscores the potential of generative AI chatbots in offering personalised, empathetic-seeming support, thereby posing considerations for traditional mental health service delivery models. It suggests that such platforms can broaden access to mental health support, particularly for younger populations, by mitigating barriers of cost, perceived stigma, and availability. However, it also raises important questions regarding the ethical implications, risks of emotional attachment to AI entities in therapeutic contexts, data privacy, and the potential for AI-induced bias or misinformation. Further research is needed to explore these aspects and establish efficacy and safety guidelines.