Major music labels sue Anthropic over alleged AI training theft

Anthropic faces a new federal lawsuit from major music industry players who allege the company stole thousands of songs to train its artificial intelligence models. Sony Music, Warner Chappell Music, and several other labels filed the complaint in the U.S. District Court for the Northern District of California on Friday, August 28. The suit claims that the maker of the Claude AI assistant improperly used a vast quantity of copyrighted material without permission. The legal action demands a jury trial and was first reported by Music Business Worldwide.

The plaintiffs assert that the defendants can no longer conceal what they describe as extraordinary theft. The lawsuit states that the mass infringement is now well documented. It further notes that another court in the district recently characterised Anthropic’s actions as straightforward piracy but at massive scale. This latest filing means that the publishing arms of all three major music companies have now taken legal action against the AI firm. The record labels are seeking statutory damages of up to $150,000 for each work willfully infringed. They also claim up to $25,000 for each alleged removal or alteration of copyright management information.

Anthropic has previously faced significant legal challenges regarding its use of copyrighted content. In October 2023, Universal Music Publishing Group, Concord Music Group, and ABKCO sued the company over roughly 500 songs. They filed a second suit earlier this year covering more than 20,000 works and seeking over $3 billion. BMG also sued Anthropic in March regarding 493 songs, while Round Hill Music brought a claim earlier this month. Despite these ongoing disputes, Anthropic told TechCrunch that it disagrees with the publishers’ claims and intends to defend itself robustly in court.

The company recently reached a historic $1.5 billion settlement with authors who accused it of improperly using their works to train Claude. This is believed to be the largest award in a U.S. copyright case. The broader judicial landscape on AI copyright remains divided. U.S. District Judge William Alsup in San Francisco described AI training as quintessentially transformative, arguing that copyright law seeks to advance original works rather than protect authors against competition. Conversely, U.S. District Judge Vince Chhabria, also in San Francisco, warned days later that widespread AI training could undermine the economic incentives that fuel human creative work.