Dual Illegibility, Ambiguous Loss, and Disenfranchised Grief: A Hybrid Model for Understanding AI Companion Bereavement.

As emotionally responsive AI systems are increasingly used for companionship, clinicians are encountering forms of distress that do not fit established models of bereavement. When these systems are altered, restricted, or lost, some users experience reactions resembling attachment disruption and relational loss. Here, AI companions are conversational systems supporting ongoing relational bonds; we address grief for their disruption, not AI simulations of deceased people. This paper argues that AI companion bereavement is best understood as a technologically mediated hybrid loss at the intersection of ambiguous loss and disenfranchised grief. We introduce the concept of dual illegibility to describe uncertainty both about what has been lost and whether that loss is socially recognised as legitimate. Drawing on grief theories and emerging research on human-AI relationships, we outline a preliminary clinical response model and consider implications for practice, ethics, and research in an evolving landscape of technologically mediated attachment.