Meta wins AI copyright case, judge welcomes other to bring lawsuits

Meta On Wednesday, he ruled against a group of 13 writers in a large copyright case containing the company’s lama artificial intelligence model, but clearly stated that the judge was limited to this case.

US Regional Judge Vince Chhabria claimed that the use of books to educate Meta’s large language models or LLMs was protected under the fair use of the US Copyright Law.

Lawyers representing the plaintiffs, including Sarah Silverman and Ta-nehisi Coates, claimed that Meta had violated the country’s copyright law, because the company did not want the authors to use their books for the company’s AI model.

In particular, Chhabria said that it was usually illegal to copy protected works without permission, but in this case the plaintiffs could not offer a compelling argument that the use of Books to educate Lama’s Lama. Chhabria wrote that the plaintiffs had put forward a two defective argument for their cases.

“In this record, Meta defeated the semi -hearted argument that the plaintiffs were caused by copying or threatening significant market damage.” He said. “This result can be an important tension with reality.”

The judge wrote that Meta was protected by the fair usage doctrine of the “copying of the work for a transformative purpose”.

“We appreciate today’s decision from the court.” He said. “Open source AI models strengthen transformative innovations, productivity and creativity for individuals and companies, and the fair use of the copyright material is a vital legal framework for creating this converting technology.”

Although the judge has valid arguments that Meta’s data training practice negatively affects the book market, the plaintiffs wrote that he did not do their case sufficiently.

Lawyers representing the plaintiffs did not respond to the request for comments.

Nevertheless, Chhabria recorded a few flaws in the defense of Meta, including the idea that the company and other enterprises are forbidden to “use the text that is protected as educational data without paying this right without paying it”.

“Meta means that such a decision will stop the development of LLMs and other productive AI technologies on their ways,” Chhabria said. “This is ridiculous.”

The judge left the door open to other authors to bring similar copyright cases about artificial intelligence against Meta and said, “In the great plan of things, the results of this decision are limited.”

“This is not a class action, so the decision only affects the rights of these thirteen writers – the works are not numerous others used to educate commodity models.” “And now, as it should be open, this decision does not represent the proposal that the use of materials protected with the right to educate the language models of Meta is legal.”

In addition, Chhabria stated that the plaintiffs that the plaintiffs may have “distributed his work illegally (through torrent)” may have been a separate claim.

At the beginning of this week, a Federal Judge decided that the use of books to educate Claude’s AI model Claude was also a “transformative” and thus satisfied the doctrine of fair use. Nevertheless, this judge said that he should make a trial on the allegations that the anthropic has sent down millions of pirated books to train AI systems. “

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: META pushes the WhatsApp ban on the devices used by the House of Representatives.