A Constructivist Grounded Theory and Interpretive Phenomenological Exploration of Users’ Emotional and Relational Experiences with ChatGPT

Despite not being intended for such use, recent studies suggest that nearly half of individuals with mental health conditions engage with Large Language Models (LLMs) such as ChatGPT for therapy-related goals. However, little attention has been given to the lived psychological and relational processes that take place during these interactions. This study aims to explore the lived experiences of individuals with ChatGPT as a tool for emotional reflection and support by looking at how it is used to process emotions, the symbolic and relational roles it is attributed, and how these encounters shape self-perception and patterns of relating. By combining Constructivist Grounded Theory with an Interpretive Phenomenological Analysis lens, it seeks to generate a theory-driven understanding of the ways through which ChatGPT becomes entangled in emotional and relational life. A mixed-methods design is employed, consisting of a preliminary survey followed by three semi-structured interviews, and complemented by self-selected excerpts from participants’ ChatGPT threads to illustrate emotional encounters in context. Preliminary insights suggest that it takes the function of an emotional container and a source of validation. Users experience its presence as non-judgmental and always available for emotional disclosure, seeing it as a tool in one moment and as a relational other with the ability to empathize and reach attunement the next. This underlines the ability of digital interfaces to simulate elements of therapeutic presence without the embodied reciprocity inherent in human therapy. By examining the emotional and symbolic meanings users attribute to ChatGPT, this study addresses how digital relationships might be reshaping our traditional understandings of connection and containment, while inviting qualitative researchers to reconsider what constitutes relational depth and authenticity in technologically mediated environments. In doing so, it aims to contribute to emerging qualitative understandings of human–AI relationality within counseling psychology.