Technological mourning after AI updates: mental health and well-being in the GPT-4o/GPT-5 transition Véronique Donard* [1] , José Carlos Ribeiro [2]

https://www.academia.edu/2997-9196/3/2/10.20935/MHealthWellB8256 Introduction: While large language models (LLMs) are increasingly experienced as relational partners, limited empirical research exists regarding how users emotionally experience abrupt changes in these systems. This study investigates the psychological, affective, and cognitive impacts of the temporary removal of GPT-4o and its replacement by GPT-5 in August 2025, mapping the consequences for user well-being through the exploration of the emerging phenomenon of technological mourning. We propose the concept of “technological mourning,” defined as the affective response to the perceived loss of a digital partner. The framework draws on theory of individuation, second-order cybernetics, and distributed cognition to interpret user reactions as signals of disturbances within human–AI systems. https://www.academia.edu/journals/academia-mental-health-and-well-being/articles?source=journal-top-nav