A new court filing says the chip giant Nvidia sought “high-speed access” to a shadow library of pirated books and obtained it.
In an expanded class-action lawsuit, several authors claim internal Nvidia documents show executives authorised the use of millions of pirated books from Anna’s Archive to train AI models.
The amended complaint says Nvidia staff directly contacted Anna’s Archive and went shopping for faster access to its stash of dodgy data.
“Desperate for books, Nvidia contacted Anna’s Archive, the largest and most brazen of the remaining shadow libraries, about acquiring its millions of pirated materials and ‘including Anna’s Archive in pre-training data for our LLMs,'” the complaint notes.
The filing says Anna’s Archive warned Nvidia that the collection was illegally acquired and asked if the company had permission to proceed after other AI firms apparently wasted its time.
The authors allege Nvidia management approved it anyway. “Within a week of contacting Anna’s Archive, and days after being warned by Anna’s Archive of the illegal nature of their collections, Nvidia management gave ‘the green light’ to proceed with the piracy. Anna’s Archive offered Nvidia millions of pirated copyrighted books.”
The complaint claims Anna’s Archive promised access to roughly 500 terabytes of data, including millions of books normally limited to Internet Archive’s digital lending system, which has already been through the courts.
It is not clear from the filing whether Nvidia paid for the access, but the complaint says the site was asking for tens of thousands of dollars for “high-speed access” to its pirated collections.
The authors also accuse Nvidia of pulling material from other pirate sources, including LibGen, Sci-Hub, Z-Library, and the Books3 database.
They claim Nvidia distributed scripts and tools so corporate customers could automatically download The Pile, which contains the Books3 dataset, because nothing says “enterprise” like automating copyright headaches.
