Recently, Authentic Brands Group, the holder of Elvis Presley's music catalog, formally filed a lawsuit against the artificial intelligence company Anthropic. The plaintiff accuses Anthropic of using 274 of Elvis's songs without permission to train its AI model, with the total damages sought reaching as high as $41 million.
The lawsuit breaks with the past practice in which such rights-protection actions were mostly initiated by the recording industry association or major record labels, becoming one of the first major legal actions brought directly against an AI development company by a well-known artist's estate. The final verdict in this case is highly likely to set a common legal precedent standard for how AI companies may lawfully use existing music catalogs in the future.
The plaintiff's lead attorney, Richard Busch, stated in the complaint that the plaintiff filed this lawsuit to firmly defend its exclusive rights, and bluntly declared that the defendant's entire business model is built upon copyright infringement on an unprecedented scale. The complaint further asserts that Anthropic's AI model, named Claude, has from its very inception been able to generate infringing content for almost every song in catalogs of all kinds across the internet.
The tracks accused of infringement include Elvis classics such as "Can't Help Falling in Love," "All Shook Up," and "A Little Less Conversation." The plaintiff demands compensation at the rate of $150,000 per infringing track, approximately 1.003 million yuan, with a total claim of roughly $41 million, equivalent to about 274 million yuan.
The complaint also specifically notes that Anthropic only set up so-called safeguards to restrict the output of infringing musical content after the matter came to light, but these protective measures were ineffective, and the company, knowing the safeguards did not work, still allowed the model to continue producing infringing music. The plaintiff believes this fully demonstrates that Anthropic acted knowingly and deliberately, with the subjective malice of continuously and intentionally infringing the copyrights of others' musical works.
At present, the underlying operating logic of AI music platforms generally involves analyzing massive amounts of recorded works to learn and master the timbre, structure, and expressive style of songs, then allowing users to quickly generate AI music via text or voice commands.
Faced with this technological trend, many musicians have repeatedly voiced their concerns publicly: on one hand, they fear their musical works and voices will be taken and used by AI models without authorization and without compensation; on the other hand, they deeply worry that shoddily made AI-generated music will ultimately completely replace excellent original songs created by humans.
In fact, some AI tools have even been openly imitating living real singers on major music streaming platforms. Just last week, the streaming giant Spotify released a set of data: over the past year, the platform has removed approximately 50 million such low-quality AI-generated tracks from its recommendation feeds.