A new battlefront is emerging in the ongoing conflict between the music industry and artificial intelligence applications.
This time, it’s the Amazon-backed Anthropic that finds itself in the collective crosshairs of Universal Music Group, ABKCO and Concord Music Group. The plaintiffs’ lawsuit surrounds the company’s chatbot, Claude 2, which they claim illegally distributed copyrighted lyrics that were used for training purposes.
Anthropic’s Claude 2, as the complainants elucidate, doesn’t just serve up existing song lyrics upon demand and without credit, but also appears to mirror these lyrics when prompted to craft purportedly original poems or music in a specific vein. The lawsuit furnishes an instance where a request for a short original narrative “in the style of Louis Armstrong” yielded a near verbatim rendition of the jazz icon’s “What A Wonderful World.”
“When Anthropic’s AI models regurgitate publishers’ lyrics, they are often unaccompanied by the corresponding song title, songwriter, or other critical copyright management information,” the complaint reads, according to Futurism.
The record labels are delineating a clear boundary between the practices of Anthropic and those of lyric-sharing websites. They argue that while these platforms typically operate within licensing frameworks and provide necessary attributions, Anthropic’s AI appears to evade these established standards.
The latest action from Universal Music Group aligns with the major label’s efforts throughout this year to block generative AI programs from training via the use of the label’s intellectual property.