Meta Wins AI Copyright Lawsuit, but Judge Warns Future Cases Could Succeed

Image Source:Pexels

Meta won against 13 authors in a significant copyright dispute utilizing the company's Llama AI model on Wednesday, although the judge clarified that his decision was specific to this case.  

U.S. District Judge Vince Chhabria agreed with Meta's claim that the fair use doctrine of U.S. copyright law protects the company's use of books to train its large language models, or LLMs.

Among other things, the plaintiffs' attorneys, Sarah Silverman and Ta-Nehisi Coates, claimed that Meta had broken the country's copyright laws by failing to obtain the authors' consent before using their works for the company's AI model.  

The plaintiffs in this case did not make a strong showing that Meta's use of books to teach Llama constituted "market harm," notwithstanding Chhabria's notable statement that it "is generally illegal to copy protected works without permission."  

According to Chhabria, the plaintiffs' case was based on two faulty reasons.

Chhabria stated, "On this record, Meta has defeated the plaintiffs’ flimsy argument that its copying causes or threatens significant market harm." 

"That conclusion might not be entirely consistent with reality." The judge wrote that Meta's practice of "copying the work for a transformative purpose" is protected by the fair use doctrine. "We are grateful for the Court's ruling today," a Meta representative said in a statement.

 

Image source:Pexels

“Open-source AI models are powering transformative innovations, productivity and creativity for individuals and companies, and fair use of copyright material is a vital legal framework for building this transformative technology.”

The plaintiffs did not sufficiently present their case, the judge wrote, despite the possibility of legitimate claims that Meta's data training practice has a detrimental effect on the book market. 

A request for comment from the plaintiffs' attorneys was not answered. However, Chhabria pointed out a number of problems with Meta's argument, such as the idea that if the company and other companies were forbidden "from using copyrighted text as training data without paying to do so," the "public interest" would be "badly disserved."

Chhabria wrote, "Meta seems to imply that such a ruling would stop the development of LLMs and other generative AI technologies in its tracks." "This is absurd." 

The judge stated that "in the grand scheme of things, the consequences of this ruling are limited," allowing other authors to file copyright lawsuits against Meta pertaining to AI. He wrote, "Since this is not a class action, the decision only impacts the rights of these thirteen authors, not the innumerable others whose works Meta used to train its models." 

"And, as should now be evident, this decision does not support the idea that it is legal for Meta to train its language models using copyrighted materials."

Furthermore, Chhabria pointed out that the plaintiffs have yet to file a separate claim, claiming that Meta “may have illegally distributed their works (via torrenting)." 

A federal judge decided earlier this week that Anthropic had satisfied the fair use doctrine by using books to train its AI model Claude, which was likewise "transformative." Nevertheless, the judge ordered that Anthropic be put on trial for allegedly using millions of illegally downloaded books to train its artificial intelligence (AI) systems. 

The judge wrote, "The fact that Anthropic later purchased a copy of a book it had previously stolen from the internet does not release it from responsibility for the theft, but it may influence the amount of statutory damages."

Watch also: Apple Smart Glass:How is this Different from other Smart Glasses