NVIDIA Contacted Anna’s Archive to Secure Access to Millions of Pirated Books

Authors suing NVIDIA for copyright infringement have filed an amended complaint alleging that the chip giant, driven by competitive pressures, contacted the "shadow library" Anna's Archive to acquire millions of pirated books for training its AI models, including NeMo and Megatron. Internal emails suggest an NVIDIA data strategy team sought "high-speed access" to the library's data, despite Anna's Archive warning about the illegal nature of its collections. The complaint claims NVIDIA management gave the "green light" to proceed, potentially accessing 500 terabytes of data, and also accuses NVIDIA of using other pirated sources like LibGen and Z-Library. Furthermore, the authors allege NVIDIA facilitated customer access to pirated datasets, leading to claims of vicarious and contributory infringement.

(Source：Torrentfreak)