US Judge Acquits "Meta" of Intellectual Property Rights Violation in Artificial Intelligence Training Case

Federal Judge Vince Chhabria ruled in favor of the company "Meta" after accusing it of using protected literary works without permission to train the artificial intelligence model "Lama," according to court documents reviewed by Agence France-Presse.

The ruling came after a group of writers, including Junot Diaz, author of "The Brief Wondrous Life of Oscar Wao," and Sarah Silverman, author of "The Bedwetter," filed a lawsuit accusing "Meta" of downloading and using their works in training the model.

However, the judge found the plaintiffs' claims "unconvincing," affirming that they did not prove Meta's violation of the "fair use" principle of content. A company spokesperson said: "We are grateful for today's decision."

Nevertheless, the judge cautioned that the ruling does not validate Meta's absolute use of protected content, warning of the risks of artificial intelligence on the publishing industry if it leads to "producing competing works that harm original creations."

On the other hand, the spokesperson for "Meta" defended open-source models, saying: "AI models enhance creativity and productivity, and reasonable use of protected content is a legal basis for developing this technology."

The ruling comes at a time when AI companies are facing a wave of lawsuits from writers and artists for using their data without permission. In a related development, a court ruled this week that the company "Anthropic" can use protected books to train its models without authors' consent.

It is worth noting that training large language models requires massive amounts of data, increasing the debate on the boundaries of "fair use" in the era of artificial intelligence.