📰 New article from TorrentFreak
NVIDIA’s Shadow Library Scripts ‘Have No Other Purpose’ Than Infringement, Judge Rules
### NVIDIA’s AI “Shortcuts” Hit a Legal Speed Bump
NVIDIA is currently the undisputed heavyweight champion of the AI boom, but it turns out even tech giants can trip over their own code. A federal judge just handed a significant blow to the chipmaker’s attempt to dodge a massive copyright lawsuit brought by a group of authors.
The drama centers on how NVIDIA trains its AI models. The plaintiffs allege that NVIDIA used datasets packed with pirated books—specifically from “shadow libraries” like Bibliotik—to fuel its learning processes.
While NVIDIA tried to argue that their software tools are just neutral technology, Judge Jon Tigar wasn’t buying it. He focused on specific scripts NVIDIA distributed to clients to help them automatically download and process these datasets. His verdict? Those scripts have “no other purpose” than to speed up copyright infringement.
A few quick takeaways from the courtroom:
- The “Inducement” Problem: The judge ruled that the claim regarding contributory infringement can move forward, specifically targeting tools tailored for piracy.
- BitTorrent is just a tool: NVIDIA tried to scrub BitTorrent references from the case, but the judge famously compared dismissing BitTorrent allegations to “asking to dismiss allegations concerning paintbrushes in a case about a dolphin painting.”
NVIDIA did manage to get one claim dismissed, but the core battle over how AI models are fed continues. This is a massive signal for the industry: if your software makes it easier to steal, you might just be liable for the theft.