Exploring Audience Perceptions of K-Pop Singers’ Vocal Identity in AI Cover Culture: Evidence from Interviews and Expert Evaluation

With the rapid development of voice cloning and singing voice conversion technologies, AI Covers have raised new questions concerning vocal authenticity, identity attribution, and audience identification.Focusing on K-pop singers' vocal identity in AI Cover culture, this study integrates grounded theory, the Delphi method, DEMATEL, and ISM.Semi-structured interviews with 16 participants generated 29 initial categories and nine principal categories, which two rounds of expert consultation refined into 26 indicators.DEMATEL and ISM were used to examine expert-perceived influence relationships and the hierarchical structure among these factors.Technological naturalness, singing-style consistency, reproduction of vocal characteristics, and emotional-expression consistency occupied relatively upstream positions, while perceived authenticity, emotional attachment, and social identification were located at the intermediate interpretive level.Identity attribution judgment was positioned closer to the outcome level and was associated with identity acceptance, negotiation, and rejection.Accordingly, the study proposes a four-stage framework: "vocal and technological perception-authenticity and emotional cognition-identity attribution judgment-vocal identity outcomes."The findings suggest that vocal identity in AI Cover culture is constructed not through vocal similarity alone, but through audiences' interpretation of vocal cues in relation to prior memories, emotional attachment, fan experience, and collective evaluation.