[JackC]

Here's the full decision, which (like most decisions!) is largely written to be legible to non-lawyers: https://storage.courtlistener.com/recap/gov.uscourts.ded.721...

The core story seems to be: Westlaw writes and owns headnotes that help lawyers find legal cases about a particular topic. Ross paid people to translate those headnotes into new text, trained an AI on the translations, and used those to make a model that helps lawyers find legal cases about a particular topic. In that specific instance the court says this plan isn't fair use. If it was fair use, one could presumably just pay people to translate headnotes directly and make a Westlaw competitor, since translating headnotes is cheaper than writing new ones. And conversely if it isn't fair use where's the harm (the court notes no copyright violation was necessary for interoperability for example) -- one can still pay people to write fresh headnotes from caselaw and create the same training set.

The court emphasizes "Because the AI landscape is changing rapidly, I note for readers that only non-generative AI is before me today." But I'm not sure "generative" is that meaningful a distinction here.

You can definitely see how AI companies will be hustling to distinguish this from "we trained on copyrighted documents, and made a general purpose AI, and then people paid to use our AI to compete with the people who owned the documents." It's not quite the same, the connection is less direct, but it's not totally different.

[anon373839]

This is an interesting opinion, but there are aspects of it that I doubt will stand the test of time.

One aspect is the court’s ruling that West’s headnotes are copyrightable even when they merely quote a court opinion verbatim, because the editorial decision to quote the material itself shows a “creative spark”. It really isn’t workable — in law specifically - for copyright to attach to the mere selection of a quote from a case to represent that case’s holding on an issue. After all, we would expect many lawyers analyzing the case independently to converge on the same quotes!

The key fact underlying all of this, I think, is that when Ross paid human annotators to write their own versions of the headnotes, they really did crib from West’s wholesale rather than doing their own independent analysis. Source text was paraphrased using curiously similar language to West’s paraphrasing. That, plus the fact that Ross was a directly competing product, is what I see as really driving this decision.

The case has very little to say about the more commonly posed question of whether copyright is infringed in large-scale language modeling.

[AnthonyMouse]

> That, plus the fact that Ross was a directly competing product, is what I see as really driving this decision.

The "competing product" thing is probably the most extreme part of this opinion.

The most important fair use factor is if the use competes with the original work, but this is generally implied to be directly competes, i.e. if you translate someone else's book from English to French and want to sell the translation, the translation is going to be in direct competition for sales to people who speak both English and French. The customer is going to use the copy claiming fair use as a direct substitute for the original work, instead of buying it.

This court is trying to extend that to anything downstream from it, which seems crazy. For example, "multiple copies for classroom use" is one of the explicit examples of fair use from the copyright statute, but schools are obviously teaching people intending to go into competition with the original author, and in general the idea that you can't read something if you ever intend to write something to sell in competition with it seems absurd and in contradiction to the common practices in reverse engineering.

But this is also a district court opinion that isn't even binding on other courts, so we'll see what happens if it gets appealed.

[mountainb]

No that is not an extreme interpretation of the fair use factors. This is a routinely emphasized factor in fair use analyses for both copyright and trademark. School fair use is different because that defense is written into the statute directly in 17 U.S.C. § 107. Also, § 108 provides extensive protections for libraries and archives that go beyond fair use doctrines.

The idea that the schools are encouraging the students to compete with the original authors of works taught in the classroom is fanciful by the meaning that courts usually apply to competition. Your example is different from this case in which Ross wanted to compete in the same market against West offering a similar service at a lower price. Another reason that the schools get a carveout is because it would make most education impractical without each school obtaining special licenses for public performance for every work referenced in the classroom.

But maybe that also provokes the question as to if schools really deserve that kind of sweetheart treatment (a massive indirect subsidy), or does it over-privileges formal schools relative to the commons at large?

[AnthonyMouse]

> School fair use is different because that defense is written into the statute directly

It's written into the statute as an example of something that would be fair use.

> The idea that the schools are encouraging the students to compete with the original authors of works taught in the classroom is fanciful by the meaning that courts usually apply to competition.

People go to art school primarily because they want to create art. People study computer science primarily because they want to write code. It's their direct intention and purpose to compete with existing works.

> Your example is different from this case in which Ross wanted to compete in the same market against West offering a similar service at a lower price.

So if you use Windows and then want to create Linux...

> Another reason that the schools get a carveout is because it would make most education impractical without each school obtaining special licenses for public performance for every work referenced in the classroom.

How is that logic any different than for AI training?

> But maybe that also provokes the question as to if schools really deserve that kind of sweetheart treatment (a massive indirect subsidy), or does it over-privileges formal schools relative to the commons at large?

It not only doesn't have any explicit requirement for a formal school (it just says "teaching"), it also isn't limited to teaching, teaching is just one of the things specified in the statute as being the kind of thing Congress intended fair use to include.

[mountainb]

>It's written into the statute as an example of something that would be fair use.

Statutory text controls what the courts can do, even and perhaps especially when it includes an example.

>People go to art school primarily because they want to create art. People study computer science primarily because they want to write code. It's their direct intention and purpose to compete with existing works.

Interesting perspective.

>So if you use Windows and then want to create Linux...

I don't understand your meaning.

>How is that logic any different than for AI training?

That is what Mark Lemley, law professor at Stanford, has argued in his many law review articles and amicus briefs: he believes that training is analogous to learning. The court here didn't agree with the Lemley view.

>It not only doesn't have any explicit requirement for a formal school (it just says "teaching"), it also isn't limited to teaching, teaching is just one of the things specified in the statute as being the kind of thing Congress intended fair use to include.

In practice courts tend to limit these exceptions to formal teaching arrangements.

[zozbot234]

Copyright covers expression, not ideas. The underlying problem here is that Ross Intelligence never went to the trouble of distilling the purely idea-based and factual element from their original sources; even their finalized search system still had a pervasive reliance on Westlaw's original and creative expression as embedded in their headnotes. Using Windows and then creating Linux is something entirely different because Linux goes to great effort in order not to use anything that's specific to Windows. Large-scale language models are probably somewhere in the middle, because their unique reliance on an incredibly wide variety of published texts makes it very unlikely that they'll ever preserve anything of substance about the expression in any single text.

[brookst]

What a world we’re in where a school using text to teach children, who will remember it, talk about it with others, likely buy it for their own children… can be framed as a “massive indirect subsidy” rather than “free advertising”.

[nerdile]

This reflects on the individuals choosing to create and proliferate such misleading or hyperbolic framing more than it does on the world that we all live in. In meatspace we usually reject these ideas and ignore the people pushing them.

[DrScientist]

The case looks pretty straightforward to me - they copied the notes ( human or machine doesn't really matter ) to directly compete with the author of the notes.

If you wrote a program that automatically rephrased an original text - something like the Encyclopaedia Britannica - to preserve the meaning but not have identical phrasing - and then sold access to that information on in a way that undercut the original - then in my view that's clearly ripping off the original creators of the Encyclopedia and would likely stop people writing new versions of the encyclopedia in the future if such activity was allowed.

These laws are there to make sure that valuable activities continue to happen and are not stopped because of theft. We need textbooks, we need journalistic articles - to get these requires people to be paid to work on them.

I think it's entirely reasonable to say that an LLM is such a program - and if used on sources which are sustained by having paid people work on them, and then the reformatted content is sold on in a way to under cut the original activity then that's a theft that's clearly damaging society.

I see LLM's as simply a different way to access the underlying content - the rules of the underlying content should still apply - ChatGPTs revenues are predicted to be in the billions this year - sending some of that to content creators, so that content continues to be produced, is not just right - it's in their interest.

[zozbot234]

> automatically rephrased an original text - something like the Encyclopaedia Britannica - to preserve the meaning but not have identical phrasing

Note that it's very hard to do this starting from a single source, because in order to be safe from any copyright concern you'd have to only preserve the bare "idea" and everything else in your text must be independent. But LLM's seem to be able to get around this by looking at many sources that are all talking about the same facts and ideas in very different ways, and then successfully generalizing "out of sample" to a different expression of the same ideas.

[DrScientist]

The concept clustering across multiple sources allows you to rephrase more accurately while retaining meaning - however the point I'm making is if you then point that program at Encyclopaedia Britannica and simply rephrase it then charge for access to the rephrased version - should you be allowed to do that?

[zozbot234]

The underlying problem is that "meaning" in the ordinary sense still includes plenty of copyrightable elements. If you point a typical LLM program at some arbitrary text and tell it to "rephrase" that, you'll generally end up with a very close paraphrase that still leaves intact to a huge extent the "structure, sequence and organization" (in a loose sense) of the original. So it turns out that you're still in breach of copyright. All you're allowed to use when starting from a single copywritten text is the ideas and facts in their very barest sense.

[bryanrasmussen]

I think that someone taking Biology 101 and ending up writing textbooks, as opposed to all the other people who just forgot what they learned once the elective was over or ended up working biologists with labs or teachers of biology and so forth, is quite different than someone saying hey I want to make a competing product to this successful company, let's take their content, re-write and use AI to make a competitor, and then actually going into direct competition with that company a couple years later

[pigbearpig]

" court’s ruling that West’s headnotes are copyrightable even when they merely quote a court opinion verbatim"

That is the opposite of the ruling. The judge said the ones that summarize and pick out the important parts are copyrightable and specifically excludes the headnotes that quote court opinion verbatim.

The judge:

"But I am still not granting summary judgment on any headnotes that are verbatim copies of the case opinion (for reasons that I explain below)"

[anon373839]

You're right as far as the MSJ is concerned, and I should've been more precise. I was focusing on the dictum in the preceding paragraph (because we're discussing the broader implications of the order rather than the nuts-and-bolts of the instant motion). In that paragraph, the judge wrote:

> More than that, each headnote is an individual, copyrightable work. That became clear to me once I analogized the lawyer’s editorial judgment to that of a sculptor. A block of raw marble, like a judicial opinion, is not copyrightable. Yet a sculptor creates a sculpture by choosing what to cut away and what to leave in place. That sculpture is copyrightable. 17 U.S.C. §102(a)(5). So too, even a headnote taken verbatim from an opinion is a carefully chosen fraction of the whole. Identifying which words matter and chiseling away the surrounding mass expresses the editor’s idea about what the important point of law from the opinion is. That editorial expression has enough “creative spark” to be original. ... So all headnotes, even any that quote judicial opinions verbatim, have original value as individual works.

I personally don't think this sculpture metaphor works for verbatim quotes from judicial opinions.

[Matticus_Rex]

Yeah, I'm willing to bet that metaphor gets called out as ludicrous by a higher court, as it has broader implications across types of editorial expression that break down when examined.

The marble from which a sculpture is carved is not itself a copyrighted work, and if we imagine it as having copyright protection, to the extent it's recognizable after editorial expression it'd have to qualify as fair use itself.

[dragonwriter]

Both the more general premise (a work must not be an infringement of someone else’s work to be a work subject to copyright) and the more specific premise (court decisions are subject to copyright in the United States) in your argument for why verbatim selection from a court decision is not analogous, for copyright, to a sculptor carving from a block of material are wrong, though.

[zozbot234]

> Yeah, I'm willing to bet that metaphor gets called out as ludicrous by a higher court, as it has broader implications across types of editorial expression that break down when examined.

It's not ludicrous at all. Whether a work of "selection" from an existing source can be copyrightable in its own right would probably have to be judged on pretty much a case-by-case basis, but even in the context of "selecting" from a ruling there are almost certainly many cases where that work is creative and original enough that it can sensibly be protected by copyright.

[bee_rider]

> It really isn’t workable — in law specifically - for copyright to attach to the mere selection of a quote from a case to represent that case’s holding on an issue. After all, we would expect many lawyers analyzing the case independently to converge on the same quotes!

I guess it depends on how long the source is, and how long the collection of quotes is, if we’d expect multiple lawyers to converge on the same solution. I don’t think it is totally obvious, though…

I’m also not sure if that’s a generally good test. It seems great for, like, painting. But I wouldn’t be surprised if we could come up with a photography scene where most professionals would converge on the same shot…

[zozbot234]

If close paraphrase can be detected, this ought to be proof enough that some non-trivial element of creativity was involved in the original text. Because purely functional and necessary elements are not protected by copyright, even when they would otherwise be creative (this is technically known as the 'scenes à faire' case) - and surely a "quote" which is unavoidable because it factually and unquestionably is the core of the ruling would have to fall under that.

[DrScientist]

Isn't the argument that the act of selecting the right quote is the real work - and the work the copier avoided in the act of copying?

You could argue that all the words are already in the dictionary - so none of them are new, you are just quoting from the dictionary in a particular order......

The reason you have people, rather than computers interpreting the law, is you can make judgements that make sense. Fundamentally these laws are there to protect work being unfairly ripped off.

What was clearly done in this case was a rip-off which damaged the original creator - everything else is dancing on the head of a pin.

[zozbot234]

Copyright does not protect work ("sweat of the brow"), it only protects expression and creativity. Thus, whenever there is only one right expression or even a bare handful in any given context, copyright does not apply to that particular choice. By analogy, arranging words in some semi-arbitrary order can be an expressive choice, whereas using what's effectively a fixed phrase is not, even though the two might look similar and involve a comparable amount of "work".

[DrScientist]

The intention of copyright is to protect useful work.

The detail of how to do that in fair way that doesn't block other people is complex[1] - you can never cover all possibilities in a written law - that's why you have people interpreting them and making judgements. All I'm saying is the guiding light in that interpretation is copyright is there to protect the justifiable work of people in a fair way.

Somebody taking those law notes and trivially copying them to directly compete is clearly not 'fair use'.

If those notes could have been created mechanically directly from the original source - why didn't the copier do that - rather than use the competitors work?

[1] given the endless creativity of humans to game systems.

[zozbot234]

> The intention of copyright is

..."to promote the progress of science and useful arts". I don't see anything in there about rewarding 'work' irrespective of whether that work involves any kind of creativity.

> If those notes could have been created mechanically directly from the original source - why didn't the copier do that

That's actually a very good question. In practice, I do absolutely agree that the notes involve plenty of originality and creativity.

[fncypants]

I think this is the best takeaway. This case and its outcome is restricted to its facts. Most of the LLM activity today is very different than what happened here.

[singleshot_]

My experience using Westlaw Keycites at work is that they’re not primarily created by fishing a quote out of a holding, but instead by synthesizing a rule. If I want a summary, I read the Keycite; if I want a money quote, I root around in the case linked to the Keycite.

Have you seen different? I’m curious what area of law you practice and in what state, for comparison’s sake.

[anon373839]

Yeah, I'd agree that most are synthesized. But I do frequently see headnotes that are verbatim or nearly verbatim slices from the text. Just grabbing a case at random: Kearney v. Salomon Smith Barney, Inc., 39 Cal.4th 95 (2006). The 4th headnote reads:

> The federal system contemplates that individual states may adopt distinct policies to protect their own residents and generally may apply those policies to businesses that choose to conduct business within that state.

And the opinion reads:

> [T]he federal system contemplates that individual states may adopt distinct policies to protect their own residents and generally may apply those policies to businesses that choose to conduct business within that state.

[6stringmerc]

The crux is Fair Use and until lobbyists change the four factor test, AI training has an uphill battle in court. It’s a very disliked observation in this forum, but I stand by my principles on this one because the courts see it my way. Derivative works, especially by artificial means, simply fail the test miserably and that’s the truth.

[greggyb]

Collections of essays or poems are considered copyrightable. This seems analogous enough to me.

[fsckboy]

>the court’s ruling that West’s headnotes are copyrightable even when they merely quote a court opinion verbatim, because the editorial decision to quote the material itself shows a “creative spark” ... when Ross paid human annotators to write their own versions of the headnotes, they really did crib from West’s wholesale rather than doing their own independent analysis

... so it follows that it was then Ross's annotators showing the creative spark

[reissbaker]

I'll quote a longer portion of the transcript about generative AI, because I think it makes the opposite of your point:

Ross’s use is not transformative. Transformativeness is about the purpose of the use. “If an original work and a secondary use share the same or highly similar purposes, and the second use is of a commercial nature, the first factor is likely to weigh against fair use, absent some other justification for copying.” Warhol, 598 U.S. at 532–33. It weighs against fair use here. Ross’s use is not transformative because it does not have a “further purpose or different character” from Thomson Reuters’s. Id. at 529.

Ross was using Thomson Reuters’s headnotes as AI data to create a legal research tool to compete with Westlaw. It is undisputed that Ross’s AI is not generative AI (AI that writes new content itself). Rather, when a user enters a legal question, Ross spits back relevant judicial opinions that have already been written. D.I. 723 at 5. That process resembles how Westlaw uses headnotes and key numbers to return a list of cases with fitting headnotes.

I think it's quite relevant that this was not generative AI: the reason that mattered is that "transformative" use biases towards Fair Use exemptions from copyright. However, this wasn't creating new content or giving people a new way to understand the data: it was just used in a search engine, much like Westlaw provided a legal search engine. The judge is pointing out that the exact implementation details of a search engine don't grant Fair Use.

This doesn't make a ruling about generative AI, but I think it's a pretty meaningful distinction: writing new content seems much more "transformative" (in a literal sense: the old content is being used to create new content) than simply writing a similar search engine, albeit one with a better search algorithm.

[BoorishBears]

I came here to point this out, and it's especially clear if you contextualize this with the original decision from September: https://www.ded.uscourts.gov/sites/ded/files/opinions/20-613...

They were doing semantic search using embeddings/rerankers.

The point that reading both decisions together compounds is that if they had trained a model on the Bulk Memos and generated novel text instead of doing direct searches, there likely would have been enough indirection introduced to prevent a summary judgement and this would have gone to a jury as the September decision states.

In other words, from their comment:

> But I'm not sure "generative" is that meaningful a distinction here.

The judge would not seem to agree at all.

[qingcharles]

Westlaw's headnotes are primarily just snippets of the case with tags attached. They are really crappy. I hate them. Some lawyers love them.

Westlaw protects them because they are the "value add." Otherwise their business model is "take published decisions the court is legally bound to provide for free and sell it to you."

An LLM today could easily recreate the headnotes in a far superior manner from scratch with the right prompt. I don't even think hallucinations would factor in on such a small task that was well regulated, but you can always just asterisk the headnotes and put a disclaimer on them.

[Tteriffic]

Exactly. Why use the headnotes at all?

I always thought they were obviously were copyrightable. Plus they’re not close to perfect either.

[AlexCoventry]

> You can definitely see how AI companies will be hustling to distinguish this from "we trained on copyrighted documents, and made a general purpose AI, and then people paid to use our AI to compete with the people who owned the documents." It's not quite the same, the connection is less direct, but it's not totally different.

Surely creating a general-purpose AI is transformative, though? Are you anticipating that AI companies will be sued for contributory infringement, because customers are using a general-purpose AI to compete with companies which created parts of the training data?

[llamaimperative]

IMO yes. The entire purpose of copyright law is to protect the incentive to create new material. A huge portion of the value prop of AI is that it captures the incentive normally bound for the creators of the training material (i.e. the whole point is you can ask the AI and not even see, never mind pay, the originator).

[zozbot234]

Ask the AI for what exactly? Factual information? That gets very low protection from a copyright point of view, especially when separate random answers by the AI will routinely show completely different rephrasings of the AI's response - implying that it can generalize well beyond the "expression" contained in any single answer, and effectively reference the underlying facts.

[AlexCoventry]

I'm not a lawyer, but I think the bar for contributory infringement is much higher than that. I think you'd have to find representatives of the defendants actually indicating somehow that people should use it that way. It seems to me that Grokster, etc.'s encouragement of their users to infringe copyright was an important factor in them losing this case, for instance.

https://supreme.justia.com/cases/federal/us/545/913/

[llamaimperative]

Encouragement is definitely not a required element. https://www.cantorcolburn.com/news-newsletters-387.html

[Ajedi32]

Interestingly, almost the entirety of the judge's opinion seems to be focused on the question of whether the translated notes are subject to copyright. It seems to completely ignore the question of whether training an AI on copyrighted material constitutes making a copy of that work in the first place. Am I missing something?

The judge does note that no copyrighted material was distributed to users, because the AI doesn't output that information:

> There is no factual dispute: Ross’s output to an end user does not include a West headnote. What matters is not “the amount and substantiality of the portion used in making a copy, but rather the amount and substantiality of what is thereby made accessible to a public for which it may serve as a competing substitute.” Authors Guild, 804 F.3d at 222 (internal quotation marks omitted). Because Ross did not make West headnotes available to the public, Ross benefits from factor three.

But he only does so as part of an analysis of whether there's a valid fair use defense for Ross's copying of the head notes, ignoring the obvious (to me) point that if no copyrighted material was distributed to end users, how can this even be a violation of copyright in the first place?

[unyttigfjelltol]

Ross evidently copied and used the text himself. It's like Ross creating an unauthorized volume of West's books, perhaps with a twist.

Obscurity ≠ legal compliance.

[Ajedi32]

So the use of AI actually has nothing to do with the ruling here? This is just about the fact that Ross made one local copy of the notes and never distributed it?

[brookst]

How would training on copyrighted material be infringement in a way that merely producing the training material (but not iterating through training) would not be?

[kevin_thibedeau]

There were data brokers who literally paid people to transcribe phone books before OCR was a viable option. That was protected, as data isn't copyrightable. It isn't hard to argue that case law metadata is no different even though it includes textual descriptions (themselves taken from public documents).

[musicale]

> "we trained on copyrighted documents, and made a general purpose AI, and then people paid to use our AI to compete with the people who owned the documents"

This is a good distillation. A bit like "we trained our system on various works of art and music, and now it is being sold as a service that competes with the original artists and musicians."

[bsder]

AI has yet to demonstrate that it can do anything different from what a group of people could sit down and do. Sure, the AI may be able to do it faster, but there hasn't yet been anything demonstrated that exceeds what humans can do.

If it would be illegal for a group of people to do something, it is also going to be illegal for an AI do so.

Why is that so surprising?

[echelon]

If the copyright holders win, the model giants will just license.

This effectively kills open source, which can't afford to license and won't be able to sublicense training data.

This is very bad for democratized access to and development of AI.

The giants will probably want this. The giants were already purchasing legacy media content enterprises (Amazon and MGM, etc.), so this will probably further consolidation and create extreme barriers to entry.

If I were OpenAI, I'd probably be very happy right now. If I were a recent batch YC AI company, I'd be mortified.

[dkjaudyeqooe]

License what? Every available copyrighted work? Even getting a tiny fraction is not practical.

To the contrary, this just means companies can't make money from these models.

Those using models for research and personal use wouldn't be infringing under the fair use tests.

[marcus0x62]

> License what? Every available copyrighted work? Even getting a tiny fraction is not practical.

Maybe the strategy is something like this:

1) Survive long enough/get enough users that killing the generative AI industry is politically infeasible.

2) Negotiate a compromise similar to the compulsory mechanical royalty system used in the music business to “compensate” the rights holders whose content is used to train the models

The biggest AI companies could even run the enforcement cartels ala BMI/ASCAP to compute and collect royalties owed.

If you take this to its logical conclusion, the AI companies wouldn’t have to pre-license anything, and would just pay out all the royalties to the biggest rights holders (more or less what happens in the music industry) on the basis that figuring out what IP went into what model output is just too hard, so instead they just agree to distribute it to whomever is on the New York Times best seller list at any given moment.

[dijksterhuis]

> the basis that figuring out what IP went into what model output is just too hard, so instead they just agree to distribute it to whomever is on the New York Times best seller list at any given moment.

the long tail exists, and there will always be a threshold for payments due to rights holders.

it used to be (like 10 years ago so i might not remember the details exactly) that if you earned less than £1 from youtube performing music rights in a quarter then any money you earned was put back into the pot and redistributed to those earning over £1.

it just wasn’t worth the cost to keep track of £0.00001 earnings for all the rights holder in the bottom of the long tail each quarter, or to pay the bank fees when the eventually earn £0.01 that can be paid to them.

definitely not perfect, but at least some people were getting paid, instead of none.

also, youtube’s data they gave us was fairly shit (video title, url). so that didn’t help. nor did the lack of compute/data proc infrastructure/skills. was historically a manual spreadsheet job trying to work out who to cut.

i had to do it a few times :/

edit —

> The biggest AI companies could even run the enforcement cartels ala BMI/ASCAP to compute and collect royalties owed.

what could happen, for music at least, is the same thing that happened with youtube, mashed up with live music analogies.

a licensing negotiation with BMI/ASCAP/PRS, and maybe major publishers directly if they get frustrated with the PROs. then PROs will use sampling of other revenue streams to work out what the likely popular things are for AI. then divvy up whatever the lump sum is between the most popular songs.

we used to do this for live music. i had to generate the sampled dataset in microsoft access each year and weed out the all the radio stings.

sorry for costing you a million pounds that one year ed sheeran :/

[DrillShopper]

> figuring out what IP went into what model output is just too hard

Check out this one cool trick companies found for skirting copyright restrictions.

Lawyers HATE them!

[AnthonyMouse]

> License what? Every available copyrighted work? Even getting a tiny fraction is not practical.

They don't need every copyrighted work and getting a fraction is entirely practical. They would go to some large conglomerate like Getty Images or large publishers or social media whose terms give the site a license to what you post and then the middle men would get a vig and the original authors would get peanuts if anything at all.

But in aggregate it would price out the little guy from creating a competing model, because each creator getting $3 is nothing to the creator but is real money to a small entity when there are a billion creators.

[jll29]

Applying copyright law more and more to things like software - and now to AI models - in other words, the status quo, makes little sense.

What is needed instead (I doubt politicians read HN, but someone go and tell them) is a new law that regulates training of these models if we want them to exist and be used in a legally safe way. This is needed for example because most jurisdictions have different copyright laws from one another, but software travels globally.

It would make sense to make all books available for non-commercial, perhaps even commercial R&D in AI, if society elected that to be beneficial in the same way that publishers must donate one copy of each new work to a copyright library (Library of Congress Library in the US, Oxford and Cambridge University libraries and British Library in the UK, Frankfurt and Leipzig Nationalbibliotheken for Germany etc.). Just add extra provisions that they need to send a plain text copy to the Linguistic Data Consortium (LDC), which manages datasets for NLP. Like for fair use, there can be provisions to make up for that use that happen automatically in the background (in some countries the price of photocopying machine includes a fee that gets passed on to copyright holders).

Otherwise you'll have one LLM being legal in one country but illegal in another because more than 15% from onw book were in the training data, and other messy situations.

[echoangle]

They didn’t train it on every available copyrighted work though, but on a specific set of legal questions and answers. And they did try to license them, and only did the workaround after not getting a license.

[synthetic-echo]

I think they were talking about the "model giants" like OpenAI you mentioned. Not saying they're correct, but I will concede the amount of copyrighted information someone like OpenAI would want is probably (at least) an order of magnitude more than this particular case.

[mvdtnz]

> License what? Every available copyrighted work? Even getting a tiny fraction is not practical.

Oh no. Anyway.

[mvdtnz]

Open source model builders are no more entitled to rip off content owners than anyone else. I couldn't possibly care any less if this impacts "democratized access" to bullshit generators. At least if the big boys license the content then the rightful owners get paid (and have the option to opt out).

[brookst]

The copyright lobby has really done a number on public policy. Copyright was never meant to be perpetual.

I’m good with your proposal if we also revert to the original 14 year + 14 year extension model. As it stands the 120 year copyright is so ridiculously tilted that we should not allow it to extend to veto power over technical advancements.

[Tanjreeve]

Legal arbitrage isn't a technical advancement. The technical advancement was all the stuff that goes into LLMs not the part where we feed ever more copyright into models for AICorp to make money.

[vkou]

I don't have either a data center, or every single copyrighted work in history to import as training data to train my open source model.

Whether or not OpenAI is found to be breaking the law will be utterly irrelevant to actual open AI efforts.

[JoshTriplett]

> If the copyright holders win, the model giants will just license.

No, they won't. The biggest models want to train on literally every piece of human-written text ever written. You can pay to license small subsets of that at a time. You can't pay to license all of it. And some of it won't be available to license at all, at any price.

If the copyright holders win, model trainers will have to pay attention to what they train on, rather than blithely ignoring licenses.

[simonw]

"The biggest models want to train on literally every piece of human-written text ever written"

They genuinely don't. There is a LOT of garbage text out there that they don't want. They want to train on every high quality piece of human-written text they can get their hands on (where the definition of "high quality" is a major piece of the secret sauce that makes some LLMs better than others), but that doesn't mean every piece of human-written text.

[ethbr1]

Even restricted to that narrower definition, the major commercial model companies wouldn't be able to afford to license all their high-quality human text.

OpenAI is Uber with a slightly less ethically despicable CEO.

It knows it's flaunting the spirit of copyright law -- it's just hoping it could bootstrap quickly enough to make the question irrelevant.

If every commercial AI company that couldn't prove training data provenance tomorrow was bankrupted, I wouldn't shed an ethical tear. Live by the sword, die by the sword.

[brookst]

Bold idea, requiring startups to proactively prove they have not broken the law. Should we apply it to all tech startups? Let’s see silicon startups prove they have not stolen trade secrets!

[pabs3]

Open source models can crowdsource open source training data. This was done for RNNoise for example.

[alberto-m]

> Here's the full decision, which (like most decisions!) is largely written to be legible to non-lawyers

For me (Italian) this is amazing! Most Italian judges and lawyers write in a purposely obscure fashion, as if they wanted to keep the plebs away from their holy secrets. This document instead begs to be read; some parts are more in the style of a novel than of a technical document.

[dkjaudyeqooe]

> The court emphasizes "Because the AI landscape is changing rapidly, I note for readers that only non-generative AI is before me today." But I'm not sure "generative" is that meaningful a distinction here.

Also the judge makes that statement, it looks like he misunderstands the nature of the AI system and the inherent generative elements it includes.

[echoangle]

How is the system inherently generative?

[currymj]

Generative is a technical term, meaning that a system models a full joint probability distribution.

For example, a classifier is a generative model if it models p(example, label) -- which is sufficient to also calculate p(label | example) if you want -- rather than just modeling p(label | example) alone.

Similar example in translation: a generative translation model would model p(french sentence, english sentence) -- implicitly including a language model of p(french) and p(english) in addition to allowing translation p(english | french) and p(french | english). A non-generative translation model would, for instance, only model p(french | english).

I don't exactly understand what this judge meant by "generative", it's presumably not the technical term.

[jll29]

Your definition of "generative" as a statistical term is correct.

However, and annoyingly so, recently the general public and some experts have been speaking of "generative AI" (or GenAI for short) when they talk about large language models.

This creates the following contradiction:

- large language models are called "generative AI"

- large language models are based on transformers, which are neural networks

- neural networks are discriminative models (not generative ones like Hidden Markov Models)

- discriminative models are the oppositve of generative models, mathematically

So we may say "Generative AI is based on discriminative (not generative) classifiers and regressors". [as I am also a linguist, I regret this usage came into being, but in linguistics you describe how language is used, not how it should be used in a hypothetical world.]

References

- Gen AI (Wikipedia) https://en.wikipedia.org/wiki/Generative_artificial_intellig...

- Discriminative (Conditional) Model (Wikipedia) https://en.wikipedia.org/wiki/Discriminative_model

[echoangle]

Do you have some kind of dictionary where I can find this definition? Because I don’t really understand how that can be the deciding factor of „generative“, and the wiki page for „generative AI“ also seems to use the generic „AI that creates new stuff“ meaning.

By your definition, basically every classifier with 2 inputs would be generative. If I have a classifier for the MNIST dataset and my inputs are the pixels of the image, does that make the classifier generative because the inputs aren’t independent from each other?

[currymj]

if you have an MNIST classifier that just takes in images, and spits out a probability of digits 1-9, that wouldn't necessarily be generative, if it is only capable of modeling P(which digit | all pixels).

But many other types of model would give you a joint distribution P(which digit, all pixels), so would be generative. Even if you only used it for classification.

https://en.wikipedia.org/wiki/Generative_model

I guess these days "generative" must mean "it is used to generate outputs that look like the training data".

But until recently, the meaning had to do with the information in the model, not how it's used.

[zozbot234]

You can derive the latter information (the joint distribution), given the former and a prior over "all pixels"-like data. So, the defining feature of "generative" models is that they feature a prior over their input data?

[Tanjreeve]

Does it really matter what the judge calls it when the ruling is about its end effects and outcomes?