Judge rules Anthropic’s training of AI with books is fair use

The use of books to train anthropic’s artificial intelligence model Claude was “fair use” and “transformative”.

Amazon-Sed Anthropic’s artificial intelligence training, big language models, “the creative elements of a particular work or even a writer’s defined style of expression,” he wrote, did not violate the copyrights of the authors.

Alsup, “To create a new text LLMs to educate the right to use copyrights to use the purpose and character of the purpose of the perfect converter.” “Like any reader who wants to be a writer.”

The decision for AI companies was an important gain as they fought legal battles on the use and implementation of copyrights in the development and training of LLMs. Alsup’s decision begins to determine the legal limits and opportunities of the industry in the future.

The anthropic spokesman said that the company was “satisfied” from the decision, and that the decision of the decision “is consistent to ensure creativity and to promote scientific progress.”

CNBC reached the plaintiffs for a comment.

The lawsuit filed at the California’s Northern Region Court of the United States was introduced by writers Andrea Bartz, Charles Graeber and Kirk Wallace Johnson in August. The case claimed that an Anthropic had “stolen hundreds of thousands of copyrights, setting up a multi -dollar business”.

Part of the case is located around the pirate and around 7 million books as part of an anthropic “central library”. The beginning decided to use these pirate materials to educate their LLMs.

Alsup ordered a trial order of how pirated books were used to create the central library of Anthropic, which will evaluate the damages.

The judge said, “This anthropic then bought a copy of a book he played on the internet, will not get rid of responsibility for theft, but may affect the degree of legal damages.”

WRISTWATCH: Anthropic explains the next AI models