"Llama AI" Used 81.7 TB of Pirated Books to Train Its Model!
Meta is accused of downloading 81.7 terabytes of pirated books to train its artificial intelligence model called "Llama." According to court documents, Meta employees obtained the data via torrents from "shadow libraries" such as Anna's Archive, Z-Library and LibGen.
M.ECONOMICTIMES.COM
A group of writers, including authors Ta-Nehisi Coates and Sarah Silverman, have filed a lawsuit against the company, alleging that Meta knowingly used the pirated books. They say internal Meta documents show that CEO Mark Zuckerberg approved the use.
REUTERS.COM
Meta argues that its use of copyrighted material to train its artificial intelligence models falls under the "fair use" doctrine. But the claims raise ethical and legal issues surrounding copyright in the development of artificial intelligence.
WIRED.COM
This lawsuit could set an important precedent for determining the legality of using copyrighted material to train AI models. Meta, like other AI companies, faces such data restrictions and legal risks.