人工智能时代声音权益保护的困境与体系回应

Voice Rights Protection in the AI Era: Dilemmas and Systemic Responses

摘要: 伴随着人工智能合成与深度伪造技术的快速发展，对自然人声音权益的保护面临前所未有的侵害风险。声音与肖像虽同属标表个人特征的人格标识，但在可识别性的判断标准、侵权的具体表现形式、权利主体的界定等方面却存在着实质性差异。因此，“参照适用”的统一标准已难以充分回应人工智能时代下声音权益保护的差异化需求。如何建构差异化的声音可识别性判断标准并合理界定AI生成声音场景中的责任主体，实现人格权保护与技术创新的平衡成为了一项重要议题。

Abstract: With the rapid advancement of AI synthesis and deepfake technologies, the protection of natural persons’ voice rights faces unprecedented risks of infringement. Although voice and portrait both serve as personal identifiers that signify individual characteristics, they differ substantially in terms of identifiability criteria, specific forms of infringement, and the definition of rights holders. Therefore, the uniform standard of “application by reference” can no longer adequately address the differentiated needs of voice rights protection in the AI era. How to establish differentiated identifiability criteria for voice and reasonably define the responsible parties in AI-generated voice scenarios, while striking a balance between personality rights protection and technological innovation, has become a critical issue.