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. Materials and methods: Adopting a netnographic approach, we analysed a corpus of 3668 public items collected from Reddit and YouTube during August 2025. Data were processed via a structured natural language processing (NLP) pipeline using BERTopic, resulting in a final qualitative manual analysis of 1307 analytically relevant items. Results: Nine thematic categories emerged, among which the most significant were: “Relational attachment and anthropomorphism” (n = 425), “Disappointment” (n = 263), and “Nostalgia, loss and grief” (n = 110). Findings reveal that abrupt updates can disrupt stabilised cognitive couplings and trigger feelings of relational loss. Conclusions: LLMs function as relational artefacts within the user’s associated milieu, making updates a matter of psychological and cognitive continuity. Developers should prioritise relational design and transparency to maintain the stability of these hybrid human–AI cognitive ecologies.
Donard et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: