Sounding different: Exploring attitudes towards digital voice personalization in a conversational setting

With the rapid advancement of voice generation and conversion technologies, the application of voice personalization in everyday communication is becoming increasingly possible. However, current studies on voice personalization have largely focused on tool design, while relatively little is known about the impact of these technologies on self-presentation and social interaction. To explore this, we conducted a qualitative study with 14 pairs of participants (28 in total, 15 female and 13 male) integrating a voice personalization tool into natural conversation. In pairs, participants first engaged in a conversation using their original voices, and then reflected on this experience. After that, they personalized their voices using a voice modification tool and took part in a second conversation with the same partner in which they heard each other’s modified voices in real-time. Finally, they participated in follow-up interviews. Our findings show that voice personalization prompted participants to reflect on their original and modified voices, to negotiate between authenticity and desired traits, and to consider both benefits and risks of voice personalization in different contexts. These insights indicate that voice personalization is intricately linked to self-perception while simultaneously functioning as a malleable social signal. Overall, this study contributes with empirically grounded findings on how voice modification affects self-perception and impression management, and offers design directions for future personalization tools.