Photo Credit: Igor Omilaev
A federal judge is reportedly poised to dismiss the majority of a copyright infringement lawsuit filed by music publishers against AI giant Anthropic.
That development emerged during a hearing yesterday on Anthropic’s motion to dismiss. From the outset, we’ve covered the defendant’s efforts to toss the complaint, which alleges that Anthropic infringed upon a number of compositions while training its Claude product and via the chatbot’s outputs.
Long story short, following one venue change and ample back and forth over the publishers’ injunction push, the plaintiffs in September urged the court to reject Amazon-backed Anthropic’s dismissal motion.
Among other things, the filing parties said they’d effectively alleged Anthropic’s “direct financial benefit from infringing activity sufficient to state a claim for vicarious liability,” intentional removal of copyright management information, and more.
Nevertheless, on the heels of the initially mentioned December 19th dismissal hearing, the court could be prepared to toss most of the suit – albeit with leave to amend.
At the time of writing, that 64-minute Zoom hearing’s transcript hadn’t yet made its way onto the docket, nor had a related ruling document. (Overall, the involved companies have jointly moved to seal a variety of materials and information in the suit.)
But according to legal trades including Bloomberg Law, the presiding judge indicated that she would likely toss two of the three claims (regarding Anthropic’s knowledge of the alleged infringement and more) for being too general.
However, the same judge reportedly found that allegations pertaining to Anthropic’s profits from the purported infringement could proceed – besides relaying that the publishers would have the chance to amend and then refile the dismissed claims.
Once again at the time of writing, the publisher plaintiffs didn’t appear to have commented publicly on the subject. As a whole, though, the news marks the latest in a line of less-than-ideal 2024 developments for rightsholders litigating against AI behemoths.
While those behemoths’ unprecedented technologies continue to evolve by leaps and bounds, infringement complaints concerning their training materials – without which the systems wouldn’t exist, by developers’ own admissions – are plodding along.
Even before this latest hiccup for music publishers, preliminary scheduling details suggested that the Anthropic case wouldn’t head to trial until 2026 at the earliest.
More immediately, regulatory battles are in full swing in several countries over generative AI training laws. One of those battles is underway in the U.K., where a government proposal would simply enable tech firms to train on protected materials. And the AI Act’s enforcement specifics remain a major focus in the EU.
