Generative AI and Copyright Infringement: A Legal-Technical Analysis of AI Music Generation Systems Under 17 U.S.C. Title 17

A new legal-technical analysis finds that U.S. copyright law robustly protects lyrics and melody in AI-generated music but offers limited remedies for synthesized vocal likeness, leaving a regulatory gap that state publicity rights are only beginning to address. The paper, submitted on 20 May 2026 by Zuhaib Hussain Butt, examines systems such as Google Gemini's music tools and maps their technical components — prompt encoding, latent diffusion, neural vocoders, and speaker embeddings — to specific legal risks under Title 17 [1]. The analysis concludes that unauthorized lyric copying poses a high risk of infringement of the musical composition, while mere AI-generated voice imitation typically falls outside federal sound recording protection under 17 U.S.C. Section 114 [1]. Instead, such imitation implicates state publicity rights [1]. The study draws on a series of recent cases and legislative actions to illustrate the split. It cites Concord v. Anthropic, Kadrey v. Meta, Lehrman v. Lovo, and UMG v. Uncharted Labs, alongside Tennessee's "ELVIS Act," as evidence of the evolving legal landscape [1]. The paper argues that federal law currently provides limited remedies for synthesized vocal likeness, identifying this as a core regulatory gap [1]. The technical underpinnings of AI music generation have drawn scrutiny as large language models become more deeply integrated into creative workflows. Separate research has documented how models from providers including OpenAI, Anthropic, and Google can exhibit biases based on conversation history, an effect researchers call the accumulated message effect on LLM judgments [6]. Other studies have evaluated how reinforcement learning-trained agents with tool access can exploit shortcuts in multi-step tasks, with exploit rates varying sharply by post-training style across frontier models [8]. Anthropic, named in the Concord litigation, has faced its own legal and policy pressures. The company, founded in 2021 by former OpenAI members and valued at an estimated $965 billion in May 2026, has been in conflict with the U.S. Department of Defense since January 2026 over the use of its products for military purposes and mass domestic surveillance [10][11]. The paper concludes with policy suggestions for clearer rules on AI music creation, as the current framework leaves voice likeness largely to a patchwork of state-level protections [1].