WHEN THE DECEASED “SPEAK”: THE ETHICS–TECHNOLOGY ENCOUNTER IN THE ERA OF POSTHUMOUS ARTIFICIAL INTELLIGENCE

The emergence of AI-driven “resurrection” systems marks a shift from static digital memorials to interactive simulations of the deceased. This paper examines the ethical and technological dimensions of posthumous identity reconstruction through griefbots and related platforms. It argues that these systems do not preserve personal identity or restore the deceased as a subject. Rather, they construct data-driven personas that simulate relational presence. The analysis addresses consent, posthumous autonomy, authenticity, data misuse, emotional effects on bereaved users, and the commercialization of grief. It also explains how large language models, retrieval-augmented generation, voice synthesis, and multimodal AI contribute to the reconstruction of communicative style and likeness. On this basis, the paper introduces a thanato-ethics of simulation: a framework for governing the asymmetry between a technologically active digital persona and the absent person it represents, centered on explicit premortem consent, transparency, human oversight, cultural sensitivity, and safeguards against emotional manipulation. The aim is not to reject technological memorialization, but to distinguish memory from simulated presence and to protect both the dignity of the deceased and the vulnerability of the living.