A federal judge in California has given preliminary approval to a $1.5 billion settlement in a major class-action lawsuit filed by authors against the AI company Anthropic. This deal is the first of its kind in a series of lawsuits targeting tech giants like OpenAI, Microsoft, and Meta over the use of copyrighted works to train AI models.
The case centers around claims that Anthropic, backed by big names such as Amazon and Alphabet, used millions of pirated books to train its AI assistant, Claude. The authors argued this was done without permission. While the judge, William Alsup, previously ruled that Anthropic’s training of the AI was fair use, he found the company violated rights by storing more than seven million pirated books in a “central library” for other potential uses.
Earlier this month, Judge Alsup had hesitated to approve the settlement and asked the parties for more information. On Thursday, during a hearing, he called the proposed agreement fair. The judge will now notify the affected authors, giving them the chance to submit claims before deciding on final approval.
The lead plaintiffs—Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson—said the judge’s decision is a step toward holding AI companies accountable and protecting creators’ rights. Maria Pallante, president of the Association of American Publishers, also praised the settlement as an important move to curb reckless use of copyrighted material by AI developers.
Anthropic’s deputy general counsel, Aparna Sridhar, welcomed the preliminary approval, saying it lets the company focus on building safe AI tools that help people and organizations alike.
A trial was planned for December to figure out how much Anthropic might owe for the alleged use of pirated books, with damages possibly reaching into the hundreds of billions. This settlement could avert that trial and set a key precedent for how AI companies handle copyrighted content in the future.