Anthropic to Pay $1.5 Billion to Authors Over Book Piracy Used to Train AI Models

The Amazon-backed tech company will hand out settlements of about $3,000 per book in the largest U.S. copyright case to date

A illustration photo shows Anthropic logo in a smartphone (Credit: CFOTO/Future Publishing via Getty Images)

Anthropic has agreed to pay $1.5 billion to settle a lawsuit from a slew of authors, who accused the AI company of illegally downloading their books to train its AI models.

The class action lawsuit is the largest payout in a U.S. copyright case, according to the plaintiff’s attorneys. Authors who opt into the agreement will be eligible for part of the $1.5 billion pot, with Anthropic paying out payments of about $3,000 per book across 500,000 authors. The settlement was reached on Aug. 26, and the plaintiff’s lawyers notified the courts on the terms of the deal on Friday.

The courts found that Anthropic, in which Amazon has a minority stake, had illegally scraped millions of books through online libraries like Library Genesis and Pirate Library Mirror in order to train their AI models with the massive amounts of text rather than purchasing the written material.

Authors Guild CEO Mary Rasenberger said the settlement deal is crucial in recognizing that authors’ creative work cannot simply be stolen.

“We expect that the settlement will lead to more licensing that gives author both compensation and control over the use of their work by AI companies, as should be the case in a functioning free market society,” she added.

The settlement comes after a previous ruling in which the court sided with Anthropic, saying that AI companies can legally train technology from purchased books not pirated ones and use it as inspiration just as any aspiring writer may purchase the book and imitate it on his or her own.

The Human Artistry Campaign, an organization which has pushed for greater accountability for AI training models, celebrated the settlement as a victory for writers, artists and creators.

“We hope this is just the first of many AI companies to be held accountable for their theft of creative work, and that it helps set a precedent that consent and compensation for works used in AI training are nonnegotiable,” the HAC stated.

Authors, who believe their work may have been stolen, can file a claim for compensation online

Comments