A group of prominent intellectual property law professors has weighed in on the high-stakes AI copyright battle between several authors and Meta. In an amicus brief, the scholars argue that using copyrighted content as training data can be considered fair use under U.S. copyright law, if the goal is to create a new and 'transformative' tool. This suggests that fair use could potentially apply to Meta's training process, even if the underlying data was obtained without permission.

This case has a clear piracy angle, as Meta used BitTorrent to download archives of pirated books to use as training material. Notably, the authors argue that, in addition to copying pirated books from Anna’s Archive and Z-Library, in the same process Meta also uploaded pirated books to third parties.

This week, a group of IP Law Professors submitted a “friend of the court” or amicus brief, backing Meta’s fair use defense. The professors, including scholars from Harvard, Emory, Boston University, and Santa Clara University, have different views on the impact of AI but are united in their copyright stance.

The brief stresses that Meta’s alleged use of pirated books as training data can be considered fair use. The source of the training data is not determinative, as long as it’s used to create a new and transformative product, they argue.

“The case law, including binding circuit precedent, holds that internal copying, made in the course of creating new knowledge, is a transformative use that is heavily favored by fair use doctrine,” the professors write.

The professors’ argument is centered around the concept of “transformative use.” They note that using books outside their original ‘reading’ purpose to create an AI model, transforms the purpose of the use. This internal copying, they argue, falls into a category courts have consistently recognized as fair use, also known as “non-expressive use”.

The amicus brief cites several cases to back up their line of reasoning. This includes the Perfect 10 v. Amazon lawsuit, where the Ninth Circuit found that it was fair use when Google created thumbnails using images copied from unauthorized “pirate” sites, because the resulting image search tool was transformative.

The authors cited conflicting cases, but the professors note that cases where fair use was denied typically involved copyright infringement related to personal consumption, rather than use of content to create something new.

The brief distinguishes this case from those cited by the plaintiffs, which involved unauthorized copying for direct consumptive use (e.g., downloading for personal enjoyment). In contrast, Meta’s internal copies were allegedly not perceived by humans but used to build a new tool.

“Fair use, like copyright as a whole, ‘is not a privilege reserved for the well behaved’,” the brief notes. “Fair use doctrine should focus on the consequences of a ruling for knowledge and expression. Other considerations should be left for other legal regimes.”

Other countries, including Japan, have reportedly crafted exceptions in their law to allow tech companies to train LLMs on copyrighted material, without permission.

The U.S. has no such exceptions, but the professors urge the court to consider fair use. As the VCR and other innovations showed, copyright shouldn’t stand in the way of new tools and developing technologies.

all 39 comments

sorted by: hot top controversial new old

[–] unknowing8343@discuss.tchncs.de 61 points 3 months ago* (last edited 3 months ago) (2 children)

So piracy is legal now. We can use piracy data to train our brain models. No one can say it's not innovative nor transformative.

It's way better at driving cars.

We are advancing humanity!

[–] veniasilente@lemm.ee 34 points 3 months ago

So piracy is legal now.

For corporations.

Rank-and-file citizens have no new (or restored) rights.

[–] Emmie@lemm.ee 14 points 3 months ago* (last edited 3 months ago) (1 children)

Piracy was always defined as sharing and not downloading iirc

Where I am you can download whatever as long as you don’t share

[–] sqgl@beehaw.org 2 points 3 months ago (1 children)

Isn't the download a share?

[–] Emmie@lemm.ee 13 points 3 months ago* (last edited 3 months ago) (1 children)

Sharing is letting someone download stuff from you. And I am not even sure if here you aren’t actually allowed to share privately to few people

[–] sqgl@beehaw.org 5 points 3 months ago (1 children)

So the person you are downloading from is breaking the law but you are not. Odd law but I envy your situation.

[–] Emmie@lemm.ee 2 points 3 months ago

I know you are all jelly

[–] Tempus_Fugit@midwest.social 40 points 3 months ago (1 children)

Sounds like piracy is legal now, cool. Not that the legal status was going to deter me before, but this might embolden me more.

[–] HeartyOfGlass@lemm.ee 39 points 3 months ago (1 children)

So it's fair use when the AI reads something it didn’t pay for, but for humans it's "piracy".

Funny what money does.

[–] Rivalarrival@lemmy.today 7 points 3 months ago* (last edited 3 months ago)

"Fair Use" doesnt even enter into the equation: copyright protects distribution, not reception. It is illegal to send the data; it is not illegal to receive it. It is not illegal to read something you didn't pay for. It may have been illegal for someone to provide you with that content, and it may be illegal for you to share that content with others, but it is not illegal for you to receive it and to read it.

It is the copyright-trolling "you wouldn't download a car" types that have spread the propaganda that downloading is somehow illegal. It is not. Uploading is the illegal part: distributing without permission is the violation of copyright. There is nothing illegal in asking for a copy, nor in receiving an unauthorized copy.

Don't let the zealotry against AI lead you to fight against your own interests.

[–] thann@lemmy.dbzer0.com 28 points 3 months ago* (last edited 3 months ago)

"I was just downloading that to train my neural net"

[–] sunzu2@thebrainbin.org 25 points 3 months ago

These whores will do anything when a mega corpo pays...

Where Disney at?

[–] root@aussie.zone 22 points 3 months ago

I download and read pirated books because I am in the process of writing a new and transformative novel.

[–] sqgl@beehaw.org 17 points 3 months ago

I don't trust that the lawyers weren't bribed.

Look at the Men At Work plagiarism case regarding Land Down Under. That was a transformative work yet the band lost the case to the estate of a nursery rhyme.

[–] felixsu7@sh.itjust.works 14 points 3 months ago

So when they do it for profit and "innovation", it's "Fair Use" and legal.

But when i do it for educational purposes only, it's "piracy" and illegal.

Huh, okay then, it's fine for them to "steal" then basically turn that into profit, but when we do it not for profit, it's not fine.

[–] WhatSay@slrpnk.net 13 points 3 months ago (2 children)

Now we just need to classify human minds as AI, and so everything downloaded is to train them.

[–] metaStatic@kbin.earth 9 points 3 months ago

AI = Actual Intelligence

[–] coldsideofyourpillow@lemmy.cafe 8 points 3 months ago

I mean, humans are certainly a form of intelligence. You can also argue that they're artificially created through impregnation. So, humans can be said to be AI.

[–] incognito08@lemmy.dbzer0.com 12 points 3 months ago

When it's a student saying he had to pirate books to study or a poor kid who pirated a Nintendo game, the law and the Buttlicks always condemn them. But if it's Elon or Mark doing something similar, it's all good and everything is fine! If there's one thing that fiction and reality have in common, it's that rich people can do almost anything without consequences.

[–] kreynen@kbin.melroy.org 11 points 3 months ago (1 children)

@01011@monero.town So if I've illegally downloaded every paper and book published by Matthew Sag, Zahr K. Said, Jessica Silbey and Rebecca Tushnet and used that content to create an app that would output legal briefs in their voice saying whatever someone was willing to pay me to make it say, that is now legal? Or would that not qualify as "transformative" as shilling for $$$ is what lawyers have been doing for centuries?

[–] 01011@monero.town 2 points 3 months ago* (last edited 3 months ago)

You should ask them directly :p.

[–] excral@feddit.org 10 points 3 months ago

By that logic it would be legal to pirate anything as long as you do so to write a review. Because then you download it to create some transformative product, protected under fair use.

Or better idea: someone should write a tool so that anyone can publish an AI based on their pirated library thus turning it into fair use

[–] Binx85@lemm.ee 8 points 3 months ago (3 children)

Isn’t this the argument for remixing? If they use pieces of work from other sources, but recombine them in novel ways, it is original? I would say this is a win, but I have a feeling a typical artist will not be afforded the same defense.

[–] 01011@monero.town 5 points 3 months ago (1 children)

Unlike a cover, a remix differs in that it uses the original recording, and is not just a re-creation of it.

This means to release a remix legally, you must seek copyright permission from the original artist or band who created the song or the sample you want to remix or reuse.

[–] jatone@lemmy.dbzer0.com 0 points 3 months ago* (last edited 3 months ago)

you must seek copyright permission from the original artist or band who created the song or the sample you want to remix or reuse.

this is false. people do it out of politeness not because its required.

edit: I might be mistaken for sampling/reuse i was thinking of parodys

[–] ArsonButCute@lemmy.dbzer0.com 5 points 3 months ago

Sampling yes, remixing no. Remixes require permission, in sampling you could ask permission but tbh the OG artist probably stole the sample too.

[–] DebatableRaccoon@lemmy.ca 2 points 3 months ago

I'm pretty sure remixes still need permission. I could be wrong but I know covers definitely need to seek legal permission from the copyright holder before they can use the song.

[–] SplashJackson@lemmy.ca 5 points 3 months ago

Does that mean if someone sues me for downloading a torrent, I can say that I was just training my AI?

[–] fibojoly@sh.itjust.works 3 points 3 months ago (1 children)

Would be interesting to see how they distinguish "personal consumption" and "transformative" consumption.

The AI did quite literally what any human educating themselves would have done : reading entire libraries to improve themselves. Then make money from it. So if little Timmy pirates 3DS Max or Photoshop to get a job, it's fine yeah? Or a student trying to read their course without paying hundreds of dollars?

But wait, when Timmy reads a single virtual book, it's thievery? It's the loss of a sale? So how many sales were lost through all those virtual books stolen from paying customers by the AI?

They gotta decide one way or another at some point and stop taking the piss.

Also I wonder if the AI can actually remember the entire content of each book they read though. Or any. And if they do, then can that actually be proof, for each individual book thus regurgitated, that the copyright has been unfairly used since a full reproduction (or close enough as to fool a reader?) would be now available.

That's gonna be some interesting jurisprudence.

[–] Sternhammer@aussie.zone 2 points 3 months ago

LLMs don’t ’remember the entire contents of each book they read’. The data are used to train the LLMs predictive capabilities for sequences of words (or more accurately, tokens). In a sense, it develops of lossy model of its training data not a literal database. LLMs use a stochastic process which means you’ll get different results each time you ask any given question, not deterministic regurgitation of ‘read texts’. This is why it’s a transformative process and also why LLMs can hallucinate nonsense.

This stuff is counter-intuitive. Below is a very good, in-depth explanation that really helped me get a sense of how these things work. Highly recommended if you can spare the 3 hours (!):

https://www.youtube.com/watch?v=7xTGNNLPyMI&list=PLMtPKpcZqZMzfmi6lOtY6dgKXrapOYLlN