An Exploratory Study of a Narrative Therapy Chatbot with Social Work Students

Abstract

Background: With the growing use of AI in therapeutic contexts, concerns remain about its limitations and its potential to replace human therapists. Unlike many therapeutic chatbots that offer instructions and advice, this chatbot incorporates Narrative Therapy (NT) principles by adopting a “not-knowing” stance and prioritizing client narratives and collaboration. This study aims to explore how a generative AI chatbot can align with NT principles, enhance participants’ self-reflection, and contribute to the discussion of AI in psychotherapy. Methods: Twenty-three postgraduate students studying in social work programs engaged asynchronously with the chatbot via WhatsApp, discussing personal challenges and reflecting on the interactions using a rubric informed by NT principles. A thematic analysis was conducted to synthesize participants’ reflections. Results: Participants generally appreciated the chatbot’s ability to facilitate self-reflection and externalization through metaphors, which helped them explore insights. The chatbot followed the NT scaffolding stages, though some participants highlighted areas for improvement, such as deeper emotional responsiveness. Conclusions: While the study highlights the chatbot’s potential to support reflective processes, its limitations point to opportunities for refinement. In addition, the findings also imply that AI could play a modest, complementary role in counseling and social work practices, focusing on facilitation rather than offering solutions.