Sony Music and Warner Music Group have initiated legal action against Anthropic, claiming the company improperly used copyrighted songs in training datasets for Claude without permission or licensing. The lawsuits allege that Anthropic engaged in large-scale scraping of copyrighted musical content—including lyrics, metadata, and audio descriptions—to enhance Claude's training data. This marks an escalation in the growing wave of copyright litigation targeting major AI companies, following similar actions against OpenAI and Stability AI by the New York Times, Getty Images, and other content creators. The record labels are seeking damages for what they characterize as systematic infringement and unauthorized commercial exploitation of protected works.
The claims center on Anthropic's data sourcing practices and raise fundamental questions about what constitutes fair use in AI model training. Unlike text-based copyright suits, which have focused on whether transformer models memorize or reproduce training data verbatim, the music industry's position emphasizes that mere incorporation of copyrighted content into training sets represents infringement regardless of output behavior. Anthropic has previously stated it sources training data responsibly and respects intellectual property rights, though the company has not yet publicly detailed its specific approach to musical content. The legal theories advanced by Sony and Warner may differ from those in the Getty and Optum cases, potentially establishing new precedent for how copyright law applies to multimodal AI systems that process audio and textual representations of creative works.
This litigation arrives as Anthropic faces heightened regulatory scrutiny across multiple jurisdictions. Notably, the EU's recent enforcement actions targeted OpenAI, ChatGPT, and other AI systems but conspicuously excluded Claude from its initial compliance sweep—a distinction that may reflect either different data practices or regulatory assessment gaps. The company's positioning as a safety-focused alternative to competitors has emphasized responsible development, making copyright allegations particularly significant to its brand positioning and investor confidence. The outcome could influence how AI companies approach training data curation and licensing agreements, potentially reshaping industry practices around content sourcing for foundation models.