Dario Amodi, Anthropic CEO, speaking on the CNBC squalk box outside the World Economic Forum in Davos, Switzerland on 21 January 2025.
Garry Miller | CNBC
anthropicUse of books to train it artificial intelligence The model cloud was “fair use” and “transformative”, a federal judge on Monday pronounced a late verdict.
Heroic-Bac anthropic’s AI training did not violate the copyright of authors Big language model American District Judge William Alsup wrote, “The creative elements of a given work have been re -introduced to the public, nor is the identified expressive expression of a writer.”
Alsup wrote, “The purpose and character of copyright work to train LLM to generate new lessons was quintextally transformative.” “There is a desire to be a writer like any reader.”
This decision was an important win for AI companies as the LLM development and training play a legal battle on the use and application of copyright functions. The decision of Alsup begins to carry forward legal limitations and opportunities for the industry.
A spokesperson of the anthropic said in a statement that the company was “pleased” with the ruling and the decision, “was in line with the purpose of copyright in enabling creativity and promoting scientific progress.”
The CNBC has reached the plaintiff for comment.
The trial filed in the US district court for the northern district of California was brought in August by writers Andrea Bartz, Charles Graber and Kirkas Johnson. The suit alleged that Anthropic stole hundreds of thousands of copyright books and “made a business of multibillion-dollars”.
A portion of the lawsuit centers is around a set of about 7 million books that are maintained as anthropic pirated and part of the “Central Library”. The startup eventually decided against using these pirated materials to train its LLM.
ALSUP ordered a test how pirated books were used to create an anthropic central library, which would evaluate any resulting losses.
The judge wrote, “The anthropic later bought a copy of a book that stolen from the Internet, it would not absent liability for theft, but it could affect the limit of legal loss,” the judge wrote.
