Meta Sued For Alleged Use Of Steel Books To Train AI Llama Model

JAKARTA - Writer of Ta-Nehision Coates, comedian Sarah Silverman and other writers, sued Meta on charges of using pirated versions of copyrighted books to train his artificial intelligence models.

In a lawsuit filed with California's federal court on Wednesday, January 8, the use of the pirated books was even carried out with the knowledge and approval of Meta CEO Mark Zuckerberg.

The lawsuit also claims that Meta leveraged datasets from LibGen, known to contain millions of pirated works, to develop their big language model, Llama.

They also said that despite internal concerns from the AI Meta executive team regarding the legality of LibGen use, Meta continued under Zuckerberg's direction.

US District Judge Vince Chhabria last year rejected claims that the text generated by Meta chatbot violated authors' copyright.

However, in a lawsuit filed on Wednesday January 8, the authors provided new evidence of a discovery process that reinforces their claim that Meta is consciously using pirated content.

Chhabria said in a hearing Thursday that she would allow writers to file amended complaints but expressed skepticism about the substance of the fraud and copyright management claims (CMI) of their books.

Earlier, in September 2024, Reuters reported that a judge ruled that the authors could examine Meta CEO Mark Zuckerberg in this AI-related copyright infringement case.

This decision is based on evidence showing Zuckerberg's direct involvement in AI training decisions.