Anthropic has settled a lawsuit from authors, who accused the Amazon-backed company of illegally downloading and copying their books to teach its AI system, in among the first deals reached by creators over novel legal issues raised by the technology.
“This historic settlement will benefit all class members,” says Justin Nelson, a lawyer for the authors. “We look forward to announcing details of the settlement in the coming weeks.”
Lawyers for both sides on Tuesday informed the court of the deal, which was reached earlier this month through mediation.
The thrust of the case — and dozens of others involving AI companies — was set to be decided by one question: Are AI companies covered by fair use, the legal doctrine in intellectual property law that allows creators to build upon copyrighted works without a license? On that issue, a judge found in June that Anthropic is on solid legal ground, at least with respect to training. Authors don’t have the right to exclude the company from using their works to train its technology as long as they purchased the books, according to the June ruling. Like any reader who wants to be a writer, the AI tool learns and create entirely new works, the court reasoned.
The technology is “among the most transformative many of us will see in our lifetimes,” wrote U.S. District Judge William Alsup.
Still, Anthropic, which declined to comment, was set to face a trial over illegally downloading seven million books to create a library that was used for training. That it later purchased copies of those books it stole off the internet earlier to cover its tracks doesn’t absolve it of liability, the court concluded. The company faced massive damages stemming from the decision that could lead to Disney and Universal getting a similar payout depending on what they unearth in discovery over how Midjourney allegedly obtained copies of thousands of films that were repurposed to teach its image generator.
Alsup’s reasoning on fair use was followed by another ruling in favor of AI companies in a lawsuit from another group of authors, including Sarah Silverman, against Meta. U.S. District Judge Vince Chhabria called the company’s utilization of the books for training “highly transformative.” By learning from the authors’ works, Meta’s large language model Llama can edit emails, write skits or provide translation services, according to the order.
In most circumstances, however, the court observed that it’s illegal for companies to use copyright-protected material to train their AI systems without permission or paymen
“No matter how transformative LLM training may be, it’s hard to imagine that it can be fair use to use copyrighted books to develop a tool to make billions or trillions of dollars while enabling the creation of a potentially endless stream of competing works that could significantly harm the market for those books,” Chhabria wrote.
With the settlement, Anthropic no longer has to face what would likely be a highly publicized trial in which a jury could’ve awarded authors a substantial payout considering damages for willfull copyright infringement reach up to $150,000 per work.