238

Nvidia accused of trying to cut a deal with Anna’s Archive for high‑speed access to the massive pirated book haul (www.tomshardware.com)

submitted 3 weeks ago by FundMECFS@anarchist.nexus to c/piracy@lemmy.dbzer0.com

25 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[-] stephen01king@piefed.zip 12 points 3 weeks ago

Was there a court case where the decision was that pirated data is legally allowed to be used for LLM training?

[-] i_stole_ur_taco@lemmy.ca 21 points 3 weeks ago

It’s 2026 in the worst timeline. You don’t ask that anymore. You ask if any entity faced consequences for doing it.

[-] far_university1990@reddthat.com 11 points 3 weeks ago

https://www.cnet.com/tech/services-and-software/meta-won-its-ai-fair-use-lawsuit-but-judge-says-authors-are-likely-to-often-win-going-forward/

Meta's use of copyrighted books to trains its Llama AI was fair use, a judge ruled.

"This ruling does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful," he wrote. "It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one."

The plaintiffs focused their arguments on how Meta's AI models can reproduce exact snippets from their works and how the company's Llama models hurt their ability to license their books to AI companies. These arguments weren't as compelling in Chhabria's eyes -- he called them "clear losers" -- so he sided with Meta.

That's different from the Anthropic ruling, where Judge William Alsup focused on the "exceedingly transformative" nature of the use of the plaintiff's books in the results AI chatbots spit out. Chhabria wrote that while "there is no disputing" that the use of copyrighted material was transformative, the more urgent question was the effect AI systems had on the ecosystem as a whole.

Maybe? Not lawyer, but sound like train might fair use? And generate not?

[-] stephen01king@piefed.zip 1 points 3 weeks ago

But that judgement clearly had nothing to do with the use of pirated material, right? It might give a partial pass to the use of copyrighted material for training LLM, but it says nothing about pirating material being legal if it is used for training LLM, which the top comment was alluding to.

[-] far_university1990@reddthat.com 8 points 3 weeks ago

https://torrentfreak.com/meta-secures-bittersweet-fair-use-victory-in-ai-piracy-case-250626/

Yesterday, U.S. District Court Judge Vince Chhabria ruled on both motions, which at first sight offers a clear win for Meta. The court denied the authors’ motion to hold Meta liable for direct copyright infringement after it obtaining pirated books from shadow libraries via BitTorrent.

Did have piracy part. Just not listed on first website.

[-] stephen01king@piefed.zip 2 points 3 weeks ago

Thanks for the source. It also seems like the distribution part is not ruled on yet, so we don't know if they'll get away with pirating stuff just yet.

[-] far_university1990@reddthat.com 1 points 3 weeks ago

Yes. Apparently meta try to only leech by modify config. But also say not use facebook server/ip to mask any seed. So not sure if actually seed. Or if matter at all.

[-] petrescatraian@libranet.de 0 points 3 weeks ago

Hmmm, that got me thinking: if you selfhost, you make sure you also instal ollama or some LLM you can also self-host. You don't need to use the LLM yourself at all. Then if something goes south, and you're accused of piracy, you can just defend yourself that you used all these materials to train your own LLM. That should get you out of trouble, right?

[-] far_university1990@reddthat.com 3 points 3 weeks ago

If you billion dollar company. Probably not if individual.

[-] petrescatraian@libranet.de 2 points 3 weeks ago

@far_university1990 yes but the legal precedent has been set, lol

(/s maybe)

[-] stephen01king@piefed.zip 1 points 3 weeks ago

Again, they haven't made a decision on the distribution part of the training data, which is the one related to piracy. They only ruled that using copyrighted material for training LLM did not satisfy copyright infringement for the specific argument used by the authors against Meta.