Nvidia defends ebook pirate web sites!

[thewayne]

This is laughable insanity.

Nvidia is defending itself in a lawsuit from a bunch of authors that their works were used - without license or any form of authorization - to train up Nvidia's LLM platform. Apparently Nvidia got the data from scraping pirate ebook web sites....

Quoth the article: "Nvidia seemed to defend the shadow libraries as a valid source of information online when responding to a lawsuit from book authors over the list of data repositories that were scraped to create the Books3 dataset used to train Nvidia's AI platform NeMo.

That list includes some of the most "notorious" shadow libraries—Bibliotik, Z-Library (Z-Lib), Libgen, Sci-Hub, and Anna's Archive, authors argued. However, Nvidia hopes to invalidate authors' copyright claims partly by denying that any of these controversial websites should even be considered shadow libraries."

Copyright infringement is a pretty simple standard, which these sites clearly violate. Now, some may contain books that are out of copyright, or completely unavailable, but a bulk of their content is illegal under U.S. law. That is pretty clear. Quibbling over the definition of the term 'shadow library' is a complete waste of the court's time and isn't going to win them any points with the judge.

This is not going to work.

https://arstechnica.com/tech-policy/2024/05/nvidia-denies-pirate-e-book-sites-are-shadow-libraries-to-shut-down-lawsuit/

Comments

[disneydream06]

I hope nothing stupid happens and those authors prevail. :)
Hugs, Jon

[kathmandu]

Nvidia may also find their product isn't marketable, and is blatantly displaying their guilt. There was already one instance of an LLM trained on AO3, that started regurgitating porn of very specific kinds, very clearly traceable.

[thewayne]

Oh, that's hilarious!  I wonder if any of my wife's porn got swept up in that?

[warriorsavant]

Frankly, I would consider the whole issue of LLM training on copyrighted material to be a clear case of infringement, but even more than most of law, copyright law depends on who has the better lawyers and deeper pockets.

BTW, just realized I've been writing "copywrite" for years, even though I know it is "the right to copy" not the "copy that is written."

[silveradept]

I would not be surprised to see nVidia decide that the best way to defend their copyright infringement is to pass off the blame onto the sites themselves that did the initial infringement. They, after all, were merely scraping content that was made freely available online. Surely the blame lies with the people who shouldn't have been posting copyrighted content online?

(This will not work, but I won't be surprised if something like that is their defense.)