分析生成式AI音乐创作的版权风险，揭示歌词抄袭与仿声的法律差异。
Generative AI and Copyright Infringement: A Legal-Technical Analysis of AI Music Generation Systems Under 17 U.S.C. Title 17
- 结合法律与技术，解析用户输入歌词并模仿歌手声音生成歌曲的侵权可能
- 未经授权使用受保护歌词构成音乐作品侵权，但仿声通常不触犯联邦录音法
- 指出联邦法对旋律和歌词保护强，但对合成嗓音缺乏有效救济，适合政策制定者参考
生成式人工智能（GenAI）使用户可通过文本提示合成音乐，融合受版权保护的歌词、AI生成的旋律及模仿真实艺人声音的合成人声。本文基于美国《专利法》第17编，分析用户将某位艺术家受保护的歌词输入GenAI系统，指令其使用另一位艺术家的声线或风格，发布并商业化该歌曲是否违反17 U.S.C. 第106条专有权利。研究整合了第17编的复制权、衍生作品权、发行权，以及第114条对录音制品的有限保护，并纳入各州新兴的声音克隆立法（如田纳西州“ELVIS法案”）。结果表明，未经授权复制歌词存在高侵权风险，而仅通过AI生成的仿声通常不在联邦录音保护范围内，转而涉及州级公开权。近期案例（Concord v. Anthropic、Kadrey v. Meta、Lehrman v. Lovo、UMG v. Uncharted Labs等）凸显这一法律分歧。文章将提示编码、潜在扩散、神经声码器、说话人嵌入等技术组件与法律风险对应，识别出监管空白：联邦法对歌词与旋律提供强保护，但对合成声线相似性缺乏有效救济机制。论文最后提出明确规则的政策建议。
原文摘要 · Abstract (English)
Generative artificial intelligence (GenAI) has enabled users to synthesize music with text prompts, combining copyrighted lyrics, AI-composed melodies, and synthetic vocals that imitate real artists. This paper examines the legal and technical dimensions of AI-based music creation (e.g., Google Gemini's music tools) under U.S. copyright law. We analyze whether a user who inputs one artist's protected lyrics into a GenAI system, directs it to use another artist's voice or style, publishes the resulting song, and monetizes it violates 17 U.S.C. Section 106's exclusive rights [3]. The analysis integrates Title 17 doctrine (rights of reproduction, derivative works, distribution), 17 U.S.C. Section 114's narrow sound recording protection [4], and the new voice-cloning laws emerging at the state level [20]. We argue that unauthorized lyric copying poses a high risk of infringement of the musical composition, whereas mere AI-generated voice imitation typically falls outside federal sound recording protection and instead implicates state publicity rights [12], [13]. Recent cases and legislation (Concord v. Anthropic [10]; Kadrey v. Meta [11]; Lehrman v. Lovo [12]; Tennessee's "ELVIS Act" [20]; UMG v. Uncharted Labs [14]; etc.) illustrate this split. We map AI technical components (prompt encoding, latent diffusion, neural vocoders, speaker embeddings) to legal risks and identify a regulatory gap: federal law robustly protects lyrics and melody but currently provides limited remedies for synthesized vocal likeness [22], [23]. The paper concludes with policy suggestions for clearer rules on AI music creation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。