本研究探討創作者身份標籤（人類與AI）對音樂評價的影響，隨著生成式AI在音樂創作中的應用日增，大眾對其情感表達與創作能力仍抱持質疑，可能導致相同作品因標示不同而評價有別。透過實驗問卷將受試者隨機分派至八組，分為僅聽單一創作者音樂與同時聽AI與人類創作的組別，比較情感、偏好與技術層面的評價差異，並探討曲風（抒情與舞曲）是否調節此效應。結果顯示，AI音樂在三層面皆低於人類作品，且在混合創作者組中，抒情曲風下身份標籤的影響更為顯著。本研究揭示聽眾對AI音樂的潛在偏見，並強調創作者標示在實驗設計中的關鍵角色，對未來AI音樂接受度與設計應用具重要參考價值。;This study examines the impact of creator identity labels (human vs. AI) on music evaluation. As generative AI becomes increasingly prevalent in music creation, public skepticism toward its emotional expressiveness and creative capacity may lead to biased judgments based solely on authorship attribution. Using an experimental questionnaire, participants were randomly as-signed to one of eight groups, either exposed to music from a single creator type or both AI- and human-labeled pieces. Evaluations were compared across emotional, preference, and technical dimensions, and genre (lyrical vs. dance) was tested as a moderator. Results showed that AI-generated music received lower ratings across all dimensions, with more pronounced differ-ences in lyrical genres within the mixed-creator condition. These findings highlight audience bias against AI-generated music and underscore the importance of authorship information in ex-perimental design, offering valuable insights into the acceptance and application of AI in creative industries.
