A supergroup of rights holders who own or control Elvis Presley songs have sued Anthropic for copyright infringement, alleging that the AI company stole The King’s works to train its AI models.

“The Elvis Plaintiffs bring this action to enforce these exclusive rights against the Defendant that has built its entire business model based on copyright infringement at a scale never seen before,” states a complaint [PDF] filed last Friday in the United States District Court for the Northern District of California.

Despite owning 274 Elvis songs, the titles of which offer many apt metaphors to illustrate the plaintiffs’ grievances, the lawsuit uses a Shakesperean allusion.

“Defendant Anthropic has essentially put to the test the adage of whether monkeys and typewriters could, provided enough time, recreate Hamlet,” the complaint states. “However, rather than wait out the ‘infinite’ amount of time the infinite monkey theorem posits is necessary for said monkeys with typewriters to randomly, one letter at a time, string together that famous question of ‘To be or not to be?’, Anthropic found a workaround.”

“Rather than training its monkeys from first principles, Anthropic found mimicry (i.e. copyright infringement) to be the much easier pursuit.”

The lawsuit uses riffs that will sound familiar to anyone who has followed the many allegations that AI companies hoovered up any content they could find to train their models:

• Fair use doctrines don’t apply to all content used to train AI models;

• Anthropic should have known that some of the torrents and datasets it used for model training were either licensed (and therefore not freely usable), or stolen;

• The AI company should know it needs a license before allowing its Claude products to do things like produce the lyrics to Elvis songs when prompted to do so;

• Chatbots cost rights-holders money when users prompt them to produce create content “in the style of” artists whose works are subject to copyright.

The plaintiffs want damages and a slice of Anthropic’s profits, or $150,000 per work. The latter values damages at $41.1 million.

Anthropic has reportedly vowed to defend the matter, but has already resolved a similar case in which it was accused of training its models on books without seeking or securing permission from the authors. The AI upstart created a $1.5 billion fund to compensate writers to make that one go away.

The plaintiffs note that Anthropic has billions of cash on hand and plans a mega-float (which it announced today) so can afford to indulge in A Little Less Conversation, Don’t Be Cruel, stop being The Devil In Disguise, and Patch It Up by paying up. ®