Voice identity conversion and cloning technologies use artificial intelligence to generate the auditory likeness of a specific human talker’s vocal identity. Given the deeply personal nature of voices, the widening availability of these technologies brings both opportunities and risks for human society. This article outlines key concepts and findings from psychological research on self-voice and other-voice perception that have a bearing on the potential impacts of synthetic voice likenesses on human listeners. Additional insights from speech and language therapy, human–computer interaction, ethics, and the law are incorporated to examine the broader implications of emergent and future voice cloning technologies.

This article has three aims: it first argues that the aesthetics of graphic novels, rarely considered in Humanities dementia research, are especially suited to narratives about traumatic dementia. Second, it argues that, within the graphic narrative genre, both indirection and realism can facilitate dementia representations. Third, it argues that the realism each author uses ‘corrects’ well-meaning, idealising, dementia images aimed at challenging negative stereotypes. In this study of Sarah Leavitt’s Tangles and Dana Walrath’s Aliceheimer’s, I show that each benefits from a particular style of realism that I call, for Tangles, ‘abstract realism’, and for Aliceheimer’s ‘adapted’ or ‘fantastic’ realism. Each graphic realism style opens up for viewers the trauma of dementia for both the dementia subject herself and for those caring for her. Images move beyond stereotypes (while not idealising), furthering, via compassion, empathy and resilience, our understanding of this challenging condition so much a part of life today.