Nvidia Allegedly Sought 'High-Speed Access' To Pirated Book Library for AI Training (torrentfreak.com)
- Reference: 0180637724
- News link: https://yro.slashdot.org/story/26/01/22/1343205/nvidia-allegedly-sought-high-speed-access-to-pirated-book-library-for-ai-training
- Source link: https://torrentfreak.com/nvidia-contacted-annas-archive-to-secure-access-to-millions-of-pirated-books/
Internal documents cited in the amended complaint show Nvidia sought information about "high-speed access" to the collection, which Anna's Archive charged tens of thousands of dollars for. According to the lawsuit, Anna's Archive warned Nvidia that its library was illegally acquired and maintained, then asked if the company had internal permission to proceed. The pirate library noted it had previously wasted time on other AI companies that couldn't secure approval. Nvidia management allegedly gave "the green light" within a week.
Anna's Archive promised access to roughly 500 terabytes of data, including millions of books normally only accessible through Internet Archive's controlled digital lending system. The lawsuit also alleges Nvidia downloaded books from LibGen, Sci-Hub, and Z-Library.
[1] https://torrentfreak.com/nvidia-contacted-annas-archive-to-secure-access-to-millions-of-pirated-books/
Corporations above the law (Score:1)
It is interesting how Copyright provisions change from law that has to be followed to suggestion when they conflict with large corporations innovating.
Somehow the dialog shifts away from compliance towards competitive pressure vs China.
My thought on the matter: We should amend our laws to just allow all LLM training on any works: By everyone , so large corporations and individuals are on equal footing.
However, 75% of any revenue from LLM output should go to owners of works used, Until the model owner can
Re: (Score:2)
> It is interesting how Copyright provisions change from law that has to be followed to suggestion when they conflict with large corporations innovating.
There has long been copyright exemptions for 'fair use' - in particular 'transformative' applications where the end result doesn't resemble the source material. Essentially the fair use exemption for 'transformative' exactly matches this sort of scenario and should be a slam dunk win for the model creators.
Re: Corporations above the law (Score:2)
Something isn't "transformative" when it can literally spit out the original work, word for word. Which it absolutely can with LLMs, because that's how they work. All they are, at heart, is a non-linear mathematical regression to a dataset, in this case the corpus of human writing. You can actually look at them as a lossy compression algorithm (technically *all* mathematical models are compression algorithms, after all, since the entire goal is to construct a model that can reproduce the input data set usin
Again? (Score:2)
Didn't we just hear about this [1]3 days ago [slashdot.org]?
[1] https://yro.slashdot.org/story/26/01/19/2257241/nvidia-contacted-annas-archive-to-secure-access-to-millions-of-pirated-books
Re: (Score:2)
At least the Slashdot moderators waited 3 days before reposting it this time. I've been recent posts where there were two identical stories posted right next to each other by different moderators.
..and nothing will happen (Score:1)
The legal case wish authors/publishers vs Meta already set the horrible precedent here. These companies could have "covered their bases" by also writing a simple script to buy a physical copy (or even e-book from Amazon/Google) of every book they used for training. That way they would have legal owner ship of one copy, to read, while relying on the pirate sites for the mass input of that data they had legally obtained a copy of.
That would have opened some legal questions, BUT that judge (who I'm sure wa
Re: (Score:2)
If Meta bought all the books they trained their AI on, it seems to me they should be allowed to do that. If you've purchased a copy of a book, you're allowed to analyze it with a computer.
Publishers' right to create a service problem (Score:2)
> That's insane. You have a right NOT to sell your IP.
Under current law, you are correct. However, previous discussions on Slashdot raised questions about what benefit there is in letting government continue to enforce the copyright owner's right to withdraw a work from publication. These contrasted Disney's "vault" (sales moratorium) practice with Valve head Gabe Newell's framing of mass infringement as "almost always a service problem" on publishers' behalf.
The end result? (Score:5, Insightful)
Anna's Archive will be prosecuted, Nvidia will be allowed to use what ever it wants and won't have to face any consequences.
That is my prediction.
Re: (Score:2)
I doubt it. If there's evidence of knowingly using something that is illegally acquired that makes you liable to be a party to the suit. If the story is correct, NVIDIA is in a world of legal hurt.
Re: (Score:2)
> NVIDIA is in a world of legal hurt.
They would buy some Trump shitcoins and all the legal troubles would evaporate.
Re: (Score:2)
> Anna's Archive will be prosecuted, Nvidia will be allowed to use what ever it wants and won't have to face any consequences.
> That is my prediction.
If the AI product nVidia ultimately creates is a thief, will we know it has become self-aware when it starts talking like a pirate?
Arrr be back.
Thanks for the tip (Score:2)
I had never heard of Anna's Library, now a whole new world of books has opened up to me! Thanks to the folks bringing the lawsuit.
Crap (Score:2)
I just realize that years of watching tiktok, youtube shorts and facebook reels has deadend my brain and I can't focus on anything longer than 30 seconds. Oh well.
AI Slashdot Editors now ! (Score:3)
Eliza can do the job better.
Plot-Twist !! (Score:4, Funny)
And you thought nobody was reading books anymore.
Re: (Score:2)
Nobody does. They train the llm on the books, so they don't have to.
The plot twist here is a bit different - it is an "intellectual property" only if it is yours, otherwise it is fair game.
Re: (Score:2)
> Nobody does. They train the llm on the books, so they don't have to.
> The plot twist here is a bit different - it is an "intellectual property" only if it is yours, otherwise it is fair game.
I asked Grok if this was possible permanently stunting brain growth and capacity harming humans long term and allowing for easy manipulation and was told not to worry about it and to trust it implicitly because its answers are always correct.