Anthropic has agreed to pay at least $1.5 billion into a class action fund as part of a settlement of litigation brought by a group of book authors.
The sum, disclosed in a court filing on Friday, “will be the largest publicly reported copyright recovery in history, larger than any other copyright class action settlement or any individual copyright case litigated to final judgment,” the attorneys for the authors wrote.
The settlement also includes a provision that releases Anthropic only for its conduct up the August 25, meaning that new claims could be filed over future conduct, according to the filing. Anthropic also has agreed to destroy the datasets used in its models.
The settlement figure amounts to about $3,000 per class work, according to the filing.
Read the terms of Anthropic’s copyright settlement.
A hearing in the case is scheduled for Sept. 8.
Last month, Anthropic and the authors’ group said that they had reached a “settlement in principle” of the creators’ lawsuit.
In June, U.S. District Judge William Alsup ruled that Anthropic’s use of the books in training models was “exceedingly transformative,” one of the factors courts have used in determining whether the use of protected works without authorization was a legal “fair use.” His decision was the first major decision that weighed the fair use question in generative AI systems.
Yet Alsup also ruled that Anthropic had to face a trial on the question of whether it is liable for downloading millions of pirated books in digital form off the internet, something it had to do in order to train its models for its AI service Claude.
“That Anthropic later bought a copy of a book it earlier stole off the internet will not absolve it of liability for the theft but it may affect the extent of statutory damages,” the judge wrote.
Alsup’s summary judgment ruling came in a case brought by a group of authors, including Andrea Bartz, author of The Lost Night: A Novel, The Herd, We Were Never Here, and The Spare Room. They argued that the use of their works to train Claude violated copyright law.
A trial in the case was scheduled to start in December.
Anthropic’s deputy general counsel, Aparna Sridhar, said in a statement, “In June, the District Court issued a landmark ruling on AI development and copyright law, finding that Anthropic’s approach to training AI models constitutes fair use. Today’s settlement, if approved, will resolve the plaintiffs’ remaining legacy claims. We remain committed to developing safe AI systems that help people and organizations extend their capabilities, advance scientific discovery, and solve complex problems.”
Despite the huge pay out, Anthropic still sees the judge’s earlier ruling as a victory, as he determined that their use of copyrighted material in training models was a “fair use.” With it unlikely that Congress will pass any kind of legislation at this point to address such use, and President Donald Trump siding with tech firms, courts have been left to resolve such disputes.
Nevertheless, the judge found potential liability in the way that Anthropic obtained the material for its models, by downloading materials from online libraries. Anthropic faced the risk of prolonged litigation and losing on that point.
The settlement comes as Hollywood studios have been stepping up challenges to the use of the intellectual property in AI systems. On Thursday, Warner Bros. Discovery sued Midjourney, claiming that an AI image-generation service resulted in AI outputs of characters from Superman to Tweety Bird. WBD was joining two other legacy studios, NBCUniversal and The Walt Disney Co., in seeking court judgments against the company.
The authors’ legal team, led by Justin Nelson of Susman Godfrey, wrote in the filing that on “a per-work basis, the settlement amount is 4 times larger than $750 statutory damages amount that a jury could award and 15 times larger than the $200 amount if Anthropic were to prevail on its defense of innocent infringement.”
