The American Startup of Artificial Intelligence (IA) Anthropic has reached an agreement through which it will pay $ 1.5 billion to end a collective action presented by writers and publishers who had claimed that the company used pirate copies of its books to train its chatbot, a program designed to simulate conversations with human users.
According to information from the Associated Press Agency (AP), the agreement still needs to be approved by a judge next Monday (8).
Three writers (Andrea Bartz, Charles Graeber and Kirk Wallace Johnson) had filed a lawsuit in a court in San Francisco last year, and the action later covered other authors and publishers.
Anthropic was accused of downloading through pirate sites, such as Pirate Library Mirror, hundreds of thousands of books to train Chatbot Claude.
In the agreement announced on Friday (5), the startup agreed to pay the authors and publishers about $ 3,000 for each of the about 500,000 books covered by the agreement.
“As far as we know, it is the largest recovery of copyright ever registered,” Justin Nelson, the authors’ lawyer, told AP. “It’s the first of its kind in the Age Age.”
In a statement, Aparna Sridhar, Anthropic’s deputy general counselor, said that today’s agreement, “if approved, will resolve the perpetrators’ pending claims,” and that the company remains “committed to developing AI insurance systems that help people and organizations expand their capabilities, promote scientific discoveries and solve complex problems.”
