Meta Wins Copyright Lawsuit Over AI Book Training

News Desk

Islamabad: A U.S. judge on Wednesday delivered a legal victory to Meta in a copyright case brought by authors who alleged the company unlawfully used their work to train its Llama AI model without permission.

District Court Judge Vince Chhabria in San Francisco ruled that Meta’s use of the copyrighted material was sufficiently “transformative” to qualify as “fair use” under U.S. copyright law—marking the second major legal win for AI companies this week.

However, the judge noted an important caveat: the authors could have made a stronger case by arguing that training powerful generative AI models with copyrighted books enables the creation of competing works, potentially threatening the original authors’ market.

“No matter how transformative generative AI training may be, it’s difficult to see how it could qualify as fair use if it leads to the development of tools that generate limitless competing content while generating enormous profits,” Chhabria wrote.

Training large language models like Llama requires vast amounts of data. This need has led to numerous lawsuits from musicians, authors, visual artists, and media outlets, who accuse AI companies of using their works without authorization or compensation.

AI firms typically defend their methods by invoking fair use, claiming that the process of training on vast datasets transforms the original content and is essential for technological progress.

“We welcome the court’s decision,” a Meta spokesperson said in response to AFP. “Open-source AI models are driving innovation, productivity, and creativity. Fair use is a critical legal principle enabling the development of this transformative technology.”

In the case at hand, the authors alleged that Meta used pirated versions of their books—including Sarah Silverman’s The Bedwetter and Junot Díaz’s The Brief Wondrous Life of Oscar Wao—to train Llama, according to court filings.

Judge Chhabria clarified that his decision should not be taken as a blanket endorsement of Meta’s use of copyrighted material in AI training. Rather, it reflects the plaintiffs’ failure to present the right legal arguments or evidence to support their claims.

A broader trend?

Earlier this week, another federal judge in San Francisco, William Alsup, issued a similar ruling in favor of AI company Anthropic, which was also sued for using copyrighted books without permission to train its Claude chatbot.

Alsup found that the use of the books in training Claude was “highly transformative” and qualified as fair use. He described the technology as potentially one of the most transformative of our era and likened AI training to the human act of learning through reading.

The case was brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who alleged that Anthropic unlawfully copied their books for AI training purposes.

Still, Alsup declined to grant Anthropic broad legal immunity, ruling that its creation of a permanent digital library using millions of pirated books was not automatically protected under fair use.

These rulings signal that while courts may recognize AI training as transformative, plaintiffs must craft stronger, more targeted arguments to challenge the practice successfully.