Mark Zuckerberg, CEO of Meta, has hired three researchers from OpenAI to bolster his company’s work on artificial intelligence. This strategic move aims to guide Meta Platforms through its current AI challenges. The social-media corporation brought aboard Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai, all of whom previously held positions at OpenAI’s branch in Zurich, as per sources privy to this information. These individuals established OpenAI’s location in Zurich towards the end of last year. Prior to their time at OpenAI, they collaborated closely within Google DeepMind, which is the organization’s artificial intelligence division. A representative from OpenAI verified that the three scientists have departed from the organization. Zuckerberg has been on an intense recruiting drive To correct the company’s AI initiatives following the underwhelming reception of its most recent model, he has sometimes proposed giving researchers up to $100 million to become part o...
By Blake Brittain
- A federal judge in San Francisco issued a ruling on Wednesday siding with Meta Platforms against a collective of authors who contended that using their books without authorization to develop their AI systems violated copyright laws.
U.S. District Judge Vince Chhabria stated in his ruling that the authors failed to provide sufficient evidence demonstrating that Meta's artificial intelligence could diminish the market for their creations, thus failing to prove that the company’s actions were unlawful according to U.S. copyright law.
Chhabria also stated that utilizing copyrighted material without authorization for AI training would be illegal in "numerous situations." This stance differs from another San Francisco judge who ruled earlier this week in a different case, determining that Anthropic’s AI training constituted “fair use” regarding copyrighted content.
Chhabria stated, "This decision doesn’t mean that Meta’s utilization of copyrighted content for training their language algorithms is legal. It merely indicates that these plaintiffs presented incorrect arguments and did not build a sufficient case to back up the correct stance."
Representatives from Meta and lawyers for the authors did not promptly reply to requests for comment.
In 2023, the authors took legal action against Meta, claiming that the corporation had utilized unauthorized copies of their books to train its AI model Llama without obtaining consent or providing payment.
This legal action is among multiple copyright lawsuits filed by authors, media organizations, and various intellectual property holders against firms such as OpenAI, Microsoft, and Anthropic concerning their AI training practices.
The legal principle of fair use permits the utilization of copyrighted material without obtaining consent from the copyright holder under certain conditions. This concept serves as an important shield for technology firms.
Chhabria's determination marks the second instance in the U.S. where fair use has been considered concerning generative AI, coming after U.S. District Judge William Alsup's verdict on the matter in the Anthropic case.
AI firms contend that their systems utilize copyrighted materials fairly by analyzing them to generate new, innovative content. They also assert that having to compensate copyright owners might impede the growth of the emerging AI sector.
The copyright holders argue that AI firms illegally replicate their creations to produce rival content, posing a threat to their earnings. During a court session in May, Judge Chhabria showed understanding of this stance, a sentiment he reaffirmed on Wednesday.
The judge stated that generative artificial intelligence could potentially inundate the market with an unlimited number of images, songs, articles, and books using only a small portion of the usual time and creative effort needed for their creation.
"By using copyrighted works to train generative AI models, these companies are producing content that frequently diminishes the market for such works, thereby significantly reducing the motivation for people to continue creating through traditional methods," explained Judge Chhabria.
(Reported by Blake Brittain in Washington; Edited by David Gregorio, Alexia Garamfalvi, and Nia Williams)