Lawyers for three authors have filed a federal lawsuit accusing AI startup Antrhopic of copyright infringement for allegedly including unauthorized copies of their the company's training data. The potential class action suit, brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, was filed in the Northern District of California San Francisco, and accuses the company of using a dataset, dubbed “the Pile,” that allegedly includes a trove of pirated works to develop its Claude AI product.
“Anthropic has enjoyed enormous financial gain from its exploitation of copyrighted material,” the complaint states, noting that Anthropic projects it will generate more than $850 million of revenue in 2024. “Anthropic’s commercial gain has come at the expense of creators and rightsholders.”
The complaint continues. “Book readers typically purchase books. Anthropic did not even take that basic and insufficient step. Anthropic never sought—let alone paid for—a license to copy and exploit the protected expression contained in the copyrighted works fed into its models. Instead, Anthropic did what any teenager could tell you is illegal. It intentionally downloaded known pirated copies of books from the internet, made unlicensed copies of them, and then used those unlicensed copies to digest and analyze the copyrighted expression-all for its own commercial gain.”
While the suit notes that Anthropic has not shared details about its training corpus for Claude, the complaint asserts that Anthropic has admitted to using “the Pile,” which lawyers described as “an 800 GB+ open-source dataset created for large language model training” and allegedly includes a controversial collection called “Books3,” said to be a trove of pirated books.
Anthropic, which was founded by ex-OpenAI employees, has positioned Claude as a powerful and secure AI service “backed by uncompromising integrity.” The company is currently facing a copyright infringement lawsuit from the music industry.
The author suit joins a host of ongoing litigation filed by authors against AI companies, including two suits filed on June 28 and July 7 of last year by the Joseph Saveri Law Firm on behalf of five named plaintiffs: Mona Awad—who has since left the case—and Paul Tremblay in the first case, and Christopher Golden, Richard Kadrey, and comedian Sarah Silverman in the second. A third action was filed in September, and includes authors Michael Chabon, David Henry Hwang, Matthew Klam, Rachel Louise Snyder, and Ayelet Waldman, among others. A court has now ordered the third case to be consolidated with the first two cases in the amended complaint.
Another author lawsuit was also filed in New York against Open AI in September 2023, by the Authors Guild and a host of authors including David Baldacci, John Grisham, and Jodi Picoult. The suits all feature nearly identical claims: that AI firms are infringing copyright by using unauthorized copies of books and articles to train AI models, including copies scraped from notorious pirate sites.
As of press time, Anthropic has not commented on the suit.
