Meta found that artificial intelligence could be trained using a secret library
Meta is being sued by a group of authors who claim that the company used books from the LibGen digital library to train its Llama AI without getting the authors' consent or paying them.
The use of the content is within the bounds of "fair use," according to Meta, which refuted the accusation and claimed that its AI, Llama 3, does not duplicate the original works.
A sizable contingent of writers reportedly declared themselves "sick" over Meta's ability to use their laboriously crafted books to train artificial intelligence, according to Sky News.
One best-selling novelist once said, "It is my entire life." "It is extremely unsettling to think that someone in Silicon Valley or anywhere else is using this work to create phony AI models."
By specifically utilizing this library, called the Parallel Library, which is purported to hold millions of copies of books, the majority of which are copyrighted, Meta is accused of breaking copyright laws in the lawsuit.
In response, Meta filed a lawsuit earlier this week, arguing that downloading these books from LibGen to train its system did not violate any copyright laws. The company says it made "fair use" of the materials and that its AI does not duplicate the authors' work.
Meta commented on the Sky News article, saying, "Fair use of copyrighted material is of paramount importance in this regard. We disagree with the plaintiffs' claims, and the filing demonstrates otherwise. We will continue to vigorously defend ourselves and protect the development of GenAI for the benefit of all."