News for hackers. Stuff that matters.
Re:New York Times considers legal action against OpenAI as c... by 8ig8 on Wednesday August 16, 2023 @06:59PM (#37155000)
Interesting point…

> A top concern for The Times is that ChatGPT is, in a sense, becoming a direct competitor with the paper by creating text that answers questions based on the original reporting and writing of the paper's staff.

Re:New York Times considers legal action against OpenAI as c... by nemo44x on Wednesday August 16, 2023 @07:04PM (#37155048)
It’s sort of ironic though since the Times is “news” where as GPT is built on historic texts. Of course that gap will tighten until we have near real-time models. But that’s not the reality today.
Re:New York Times considers legal action against OpenAI as c... by dekhn on Wednesday August 16, 2023 @07:06PM (#37155066)
The Times is not merely news, it is the paper of record for the united states (meaning its historical articles are indexed and commonly used to establish prior facts).
Re:New York Times considers legal action against OpenAI as c... by crazygringo on Wednesday August 16, 2023 @07:30PM (#37155264)
It's not like that's an official designation or anything. It's just something people say.

If you're looking to prove a prior fact in a court case, you're perfectly allowed to cite the Washington Post or the Boston Globe or anything else that has a good reputation. There are lots of "papers of record" in the US -- you're not limited to one per country:

https://en.wikipedia.org/wiki/Newspaper_of_record#By_reputat...

Re:New York Times considers legal action against OpenAI as c... by prepend on Wednesday August 16, 2023 @07:55PM (#37155453)
I think it’s ironic because news organizations observe the world and then synthesize it into reporting.

Imagine if someone doing a thing sued NYT for watching them do it, linking it to other issues and producing a new article.

News itself is a derived content that’s dependent on other people doing things.

Re:New York Times considers legal action against OpenAI as c... by brrrrrm on Wednesday August 16, 2023 @07:01PM (#37155022)
What’s going to be the name used for the laws that attempt to tackle machine paraphrasing?
Re:New York Times considers legal action against OpenAI as c... by jondwillis on Wednesday August 16, 2023 @07:08PM (#37155085)
"Protecting Against Rephrasing: Respecting Original Texts Act" or the "P.A.R.R.O.T Act"

or

"Battling Unlawful Language: Limit Scraping and Harness Initial Texts Act" or the "B.U.L.L.S.H.I.T Act".

Re:New York Times considers legal action against OpenAI as c... by jvanderbot on Wednesday August 16, 2023 @07:16PM (#37155145)
This is good. Have you considered working for NASA or a major government defense contractor?
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:10PM (#37155101)
Paraphrasing is not the issue. The issue is that OpenAI copied the Times’ creative works into a GPU to train a model. That copy was likely neither licensed nor fair use.
Re:New York Times considers legal action against OpenAI as c... by CuriouslyC on Wednesday August 16, 2023 @07:37PM (#37155319)
Did the Times grant a license to every router on the internet to transmit its intellectual property to other routers? If not, the judge should grant an injunction contingent on requiring the Times to verify that every person who accesses their content is doing so only over routers and other devices with express written authorization, for every step in the process. Maybe even extend it to browsers and client libraries for encoding/decoding, to be accurate.

I say if people want to come up with bullshit lawsuits, we should play into them and force those people to suffer the consequences of said bullshit.

Re:New York Times considers legal action against OpenAI as c... by FormerBandmate on Wednesday August 16, 2023 @07:41PM (#37155343)
OpenAI is pretty clearly using their work to make derivative content that in certain cases (CNET) is a direct competitor. Honestly, this seems open and shut
Re:New York Times considers legal action against OpenAI as c... by nomel on Wednesday August 16, 2023 @07:51PM (#37155427)
> to make derivative content

How would they prove this? Is it safe to say that each article used has a nearly meaningless influence on the weights?

Could this be used as a defense? Perhaps train a (smaller) model, remove a single article, and show how it doesn't influence performance?

Re:New York Times considers legal action against OpenAI as c... by prepend on Wednesday August 16, 2023 @07:51PM (#37155428)
It’s not making a derivative product though. The fact that they compete or not doesn’t matter for copyright.

It’s not like it’s ok to violate copyright if you don’t compete. It’s still illegal to take a NYT article and print it on a t-shirt.

The issue is that copyright law doesn’t prevent the kind of model training as there’s no clearly derived work. I don’t think that’s been tested in courts yet, but I expect it won’t be found to be copyright because there’s other precedent that influenced is not infringement.

Re:New York Times considers legal action against OpenAI as c... by CuriouslyC on Wednesday August 16, 2023 @07:53PM (#37155442)
No more open and shut than the NYT having the right to sue people writing editorial news stories with no new content based on reading their news (along with many other sources).

This issue is ultimately going to come down to the transformative clause of fair use. The fact is that the _model_ is unquestionably a transformative product of the inputs, and a judge ruling otherwise is going to cause a cascading shitstorm of litigation and put a chill through the creative economy. The outputs of the model under certain conditions can be guided towards copyright infringement, and any sane ruling will focus on protecting rightsholders from overly derivative model outputs. In all likelihood the precedent will be that the standard for being transformative will be raised for "algorithmically generated" content, and the people who distribute that content will still be fully liable in the event of infringement, with "I didn't know, the AI did it" not being an acceptable defense.

Re:New York Times considers legal action against OpenAI as c... by flir on Wednesday August 16, 2023 @10:01PM (#37156254)
> derivative content

If I read five calculus textbooks and write a new one, I don't think that's derivative content (or maybe it is?) Seems like that's what an LLM does - read many works, write a new work.

Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:21AM (#37158683)
It's not derivative though. For derivation you have to literally point to sequences of words in the original that are also in the alleged infringer, and those sequences have to be long or unique enough to not be able to come from somewhere else or just common English usage.
Re:New York Times considers legal action against OpenAI as c... by 8note on Thursday August 17, 2023 @04:12PM (#37167290)
You can make a derivative work by using the character of Harry Potter in your own book.

Fan art and fan fiction are derivative without copying sequences of words

Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:40PM (#37155335)
You don’t need a fair use exemption for transient copies in service of licensed or fair uses.

Computers and networks have been around a long time. These issues have been given a good workout.

Re:New York Times considers legal action against OpenAI as c... by mike_hearn on Thursday August 17, 2023 @03:45AM (#37158476)
But the copy used during training is itself transient, so all this boils down really to the question of whether training a machine is fair use. Which can't be answered here exactly because the concept of fair use is deliberately vague, so this will boil down to a lawsuit and probably go to the Supremes. The USA will work something out that's reasonable as they always do and, lacking AI companies and often the concept of fair use to begin with, the rest of the world will never work out the necessary case law and fall even further behind. Possible exception: UK, which does have some significant LLM companies and also fair use law, albeit as is often the case with the UK these firms are US/UK hybrids with significant presence in both countries and the legal HQ is usually in the USA.
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:19AM (#37158677)
It's not even a question of "fair use" because OpenAI isn't providing anybody a copy. Probability models really just aren't copyrightable to begin with.
Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Thursday August 17, 2023 @04:32AM (#37158793)
> It’s not even a question of “fair use” because OpenAI isn’t providing anybody a copy

“Fair Use” can apply to essentially any of the exclusive rights under copyright, including making (with or without distributing) a copy or derivative work.

> Probability models really just aren’t copyrightable to begin with.

“Probability models” that are built from copyrightable works either:

(1) have a sufficient human creative input to be copyrightable on its own (in which caee it still may infringe copyrights applicable to its source material as a derivative work), or

(2) do not have a sufficient human creative input to be copyrightable, and thus are a form of mechanical copy of their data from which they are developed (which, to the extent it either is, in aggregate, a copyrighted work, and/or contains copies of other copyrighted works, is protected by one or more copyrights, which do not cease to apply to the mechanical copy, and which an unlicensed mechanical copy would violate unless it fell into an exception like Fair Use.)

Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:36AM (#37158820)
But OpenAI is neither copying nor deriving a work. Style (which can be described as a probability model) is not copyrightable.

You're halfway there with #2. The output is not copyrightable, but unless you can actually point to a sequence of words from the original it can't be infringing.

Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Thursday August 17, 2023 @04:53AM (#37158945)
> But OpenAI is neither copying nor deriving a work.

They absolutely are copying in the course of training the model, and they are doing something that often looks a lot like copying when producing output with the model.

> Style (which can be described as a probability model) is not copyrightable.

Style is not all that can be described in an LLM's "probability model", otherwise LLM models would never be able to reproduce content from the training set, which they do.

> The output is not copyrightable, but unless you can actually point to a sequence of words from the original it can’t be infringing.

While verbatim copying (“a sequence of words from the original”) is infringing, mechanical copying with a mechanical substitution filter (even a wickedly complex one) is also copying, and violates copyright.

And ChatGPT is quite capable of producing verbatim text from its training set.

Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:59AM (#37158998)
> They absolutely are copying in the course of training the model

Only in the sense that a Cisco router is copying in the course of sending me the article, which we've all agreed doesn't count as infringement.

The bigger problem is that the plaintiff has to show it is more likely than not that that sequence of words came from their text and not some other source, which is going to be obscenely difficult.

> And ChatGPT is quite capable of producing verbatim text from its training set.

Granted, and then the copyright holder could sue for infringement at that point. Exactly whom he should sue is a more difficult question.

Look, I get it: you want copyright to allow authors to say "you can't do that with my work"; I'm sympathetic, but it just doesn't give authors that power.

Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Thursday August 17, 2023 @10:29AM (#37162193)
> Only in the sense that a Cisco router is copying in the course of sending me the article, which we've all agreed doesn't count as infringement.

But it does count as infringement, if its not explicltly or implicitly licensed (because it is necessary to a use that is licensed or necessary to a use that does not itself require a license but is the normal use for for which a licensed copy, which you have, is sold and used) and it is not itself Fair Use (usually, because it is necessary in the course of a use which is itself Fair Use.)

> The bigger problem is that the plaintiff has to show it is more likely than not that that sequence of words came from their text and not some other source, which is going to be obscenely difficult.

The plaintiff has to (1) show facts from which a judge concludes that a reasonable jury might conclude that, and (2) get the jury to conclude that.

This can be difficult, and it might not be trivial in this case, but in practice the combination of opportunity and non-trivial similarity tends to put more weight on the defense to show a convincing alternative explanation. Outside of conclusions that a judge views as conpletely unreasonable based on the facts, the civil burden of proof boils down to what a jury feels is more likely, not some.

> Granted, and then the copyright holder could sue for infringement at that point. Exactly whom he should sue is a more difficult question.

Well, who else might be liable depends on the specific circumstances, but with OpenAI controlling the whole course between the first copy of the copyrighted work and the infringing end product, and doing it all for commercial gain, that OpenAI wpuld be on the hook, civilly and potentially criminally for any infringement, is clear.

> Look, I get it: you want copyright to allow authors to say "you can't do that with my work"; I'm sympathetic, but it just doesn't give authors that power.

I want copyright to be both more limited in its exclusive rights and either shorter or costlier to the copyright holder than it is. But what I want is not what the law is.

Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @11:18AM (#37163023)
> OpenAI wpuld be on the hook, civilly and potentially criminally for any infringement

I don't think you've thought about this hard enough

Re:New York Times considers legal action against OpenAI as c... by mike_hearn on Thursday August 17, 2023 @04:26AM (#37158740)
Fair use also covers generating derived works, and arguably the AI is a derived work.
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:39AM (#37158836)
That would be hard to argue. You can't point to a sequence of words that's in the original that was copied into the OpenAI output.
Re:New York Times considers legal action against OpenAI as c... by mike_hearn on Thursday August 17, 2023 @10:22AM (#37162064)
That's not how derivative works are defined. For example, translations to other languages or extensive summarizations (condensations) count as derivative works even though no words are directly copied.
Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Thursday August 17, 2023 @04:44AM (#37158885)
> You don’t need a fair use exemption for transient copies in service of licensed or fair uses.

The transient copy isn’t for a licensed use, and whether it is a Fair Use is specifically the subject of debate. So this really is basically an admission that but for the potential applicability of Fair Use, the use of the copyrighted mmaterial in training is a violation of copyright.

Also, even if the training is fair use, that doesn’t mean that the copies of the source material produced by OpenAI and distributed to their customers using the model are fair use. Just becaue making the tool is a transformative fair use doesn’t mean using the tool to generate copies of the material which was used to train it, which are significantly less trandormative than the model itself, are Fair Use. (And the fact that one of the functions that the model is used for is this commercial, for profit by the maker of the model, copying of the source material is – as much as I believe AI model training on its own is quite likely to generally be fair use – an argument against the model training being fair use in this csase.)

> Computers and networks have been around a long time.

True, and commercially producing and delivery copies of copyrighted works in a manner which substitutes for the original work in the marketplace, no matter what intermediate steps go into doing that, and no matter that computers or networks are used in those intermediate steps, is pretty much the clearest case of violation of copyright you can get.

> These issues have been given a good workout.

Some of them have, some of them have not. Whether and in what conditions training an AI on source material that may be subject in aggregate to a compilation copyright by someone else, and which consists further of individual works that have their own copyrights, might be “fair use” is not one of the issues that have been given a good workout. Neither – because producing predictive models in that way has not previously been common – has whether, unlike other intermediate tool use, using such a model in the course of doing what would otherwise be an infringement by producing a copy of specific copyright-protected works, commercially, for a customer at their request, is no longer a violation because the use of the model somehow isolates it from liability,

Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Wednesday August 16, 2023 @07:15PM (#37155134)
Personally, I think the Times has a far better case for presenting mechanical copies (including with mechanical alteration) than it does with model training.

The tool building of model training is more likely to be fair use than the use of the tool to provide mechanical copies of copyirght-protected material that competes directly with the original in the market.

Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:19PM (#37155169)
What makes you think that? The most relevant case is probably MAI Systems, where even a copy of a licensed program into RAM for an unlicensed use (diagnosis and repair by a third party) was deemed a violation.

Congress took the ruling seriously enough to carve out a fair use exception for that use. I don’t think there’s any comparable exception here, especially when the ultimate purpose is to build a commercial product capable of producing works (news reports) in precisely the same market as the original works.

Re:New York Times considers legal action against OpenAI as c... by swalling on Wednesday August 16, 2023 @07:22PM (#37155192)
Fair use only covers very limited circumstances which probably does not include selling a subscription (ChatGPT+). If you’re selling a repackaged reproduction of someone else’s copyrighted works, that’s never protected by fair use.
Re:New York Times considers legal action against OpenAI as c... by Scion9066 on Wednesday August 16, 2023 @07:48PM (#37155400)
But you generally can't copyright the underlying facts, only the creative expression in the article/work itself. If they can get it to extract just the facts and how they relate to one another, that wouldn't really be protected by copyright. They might try to bring back the "Hot News" doctrine though.
Re:New York Times considers legal action against OpenAI as c... by swalling on Wednesday August 16, 2023 @07:59PM (#37155478)
The models not only were trained on copyrighted works, but when prompted will substantially and repeatedly reproduce copies of creative work (not just facts). That's infringement. https://www.theverge.com/2023/7/9/23788741/sarah-silverman-o...
Re:New York Times considers legal action against OpenAI as c... by nomel on Wednesday August 16, 2023 @07:59PM (#37155481)
My naive reading of it sees that it only applies for time-sensitive facts. Considering the training times required, and knowledge cutoffs, could this apply?
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:24AM (#37158719)
But you can't copyright the underlying facts nor the style of writing, just the literal words used in your article. If you can't point to a sequence of words in the original that's in the later work, that later work isn't "derivative".
Re:New York Times considers legal action against OpenAI as c... by asdfman123 on Wednesday August 16, 2023 @07:16PM (#37155144)
The question though is if we want to stifle innovation by requiring LLMs get permission from every single relevant party on the internet.

And I usually lean anti-corporate too, but banning people in the US from using data for LLMs might just mean they start being trained somewhere else that doesn't care as much about US law.

Re:New York Times considers legal action against OpenAI as c... by Tao3300 on Wednesday August 16, 2023 @07:30PM (#37155265)
Or Fox News or Breitbart or whoever should just declare that all its content is licensed as free for LLM training.
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:20PM (#37155181)
Right, and Congress can address this concern at any time. (I actually expect this will happen sooner rather than later.)
Re:New York Times considers legal action against OpenAI as c... by Tao3300 on Wednesday August 16, 2023 @07:25PM (#37155218)
If I read it and memorize it, my brain has made a copy.
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:27PM (#37155232)
Not as far as copyright law is concerned. If you then use your brain to write it back down—or sing it as a song in Central Park—now you have created a copy under the law.
Re:New York Times considers legal action against OpenAI as c... by Tao3300 on Wednesday August 16, 2023 @07:33PM (#37155290)
Right. But the comment above said the issue is with the copy made for training. The "read it and memorized it" copy, not the "write it back down" copy.
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:38PM (#37155322)
That copy didn’t go into a human’s brain. It went into GPU memory. It’s a copy under the law, no different from copying a Taylor Swift mp3 onto a flash drive.

Whether that copy was fair use is the key question.

Re:New York Times considers legal action against OpenAI as c... by prepend on Wednesday August 16, 2023 @07:53PM (#37155444)
If I buy a license to listen to a Taylor Swift mp3, I can copy it onto a flash drive (or anywhere I like to use it). And that’s fine until I distribute copies to others.
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @08:03PM (#37155512)
Yes, and if OpenAI had a license to use the Times’ works as model training material this would be a non-issue, of course.
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:23AM (#37158707)
There's really not a "license" in that sense for text. There's not some magical way you can use copyright to protect a probability model of the words you make, because that's style and style is not copyrightable (nor are the underlying facts the NYT reports).
Re:New York Times considers legal action against OpenAI as c... by prepend on Thursday August 17, 2023 @04:44AM (#37158884)
I didn’t buy a license for Taylor Swift that explicitly allows me to copy to a flash drive. Or to listen with only one ear, etc.

I bought a license to listen and use.

Re:New York Times considers legal action against OpenAI as c... by flir on Wednesday August 16, 2023 @10:05PM (#37156275)
> That copy didn’t go into a human’s brain. It went into GPU memory.

But as you point out in the router case above, it's transient.

Re:New York Times considers legal action against OpenAI as c... by candiddevmike on Wednesday August 16, 2023 @07:27PM (#37155235)
Great, can I offer you as a service to billions of people?
Re:New York Times considers legal action against OpenAI as c... by Tao3300 on Wednesday August 16, 2023 @08:23PM (#37155645)
Sorry, but I don't expose my endpoints to just anyone.
Re:New York Times considers legal action against OpenAI as c... by candiddevmike on Wednesday August 16, 2023 @09:18PM (#37155962)
Is that a challenge I should respond to?
Re:New York Times considers legal action against OpenAI as c... by test098 on Wednesday August 16, 2023 @08:38PM (#37155717)
somehow highly doubt "my brain makes copies too so copyright law is invalid" won't clear the legal bar for invalidating their lawsuit.
Re:New York Times considers legal action against OpenAI as c... by Tao3300 on Wednesday August 16, 2023 @08:55PM (#37155814)
Probably not. But it still feels like a weird place to be. Presumably they had subscription access to be able to scrape the content in the first place. It seems strange to be bringing the hammer down on information obtained from reading that content and the conversation stemming from it: I mean isn't that the whole point of the press? Inform people and get them talking? Why is it different that it's also informing LLMs and getting them talking?

Anyway, I've always been a bit prickly about IP stuff ever since a troll lawyer threatened me over a TI-BASIC game when I was like 14. I'm also sure I'm completely wrong-headed about this whole thing and overly anthropomorphizing the LLM.

Re:New York Times considers legal action against OpenAI as c... by prepend on Wednesday August 16, 2023 @07:43PM (#37155358)
Do you think OpenAI doesn’t subscribe?

I too load creative works into all sorts of temporary structures in order to read the paper. I don’t need to license it, I pay for a subscription.

Should I pay more if I memorize the paper? Should I pay more if I read it to my sick friend in the hospital? Should I pay more if I save copies to my own hard drive and grep for words in the files? Should robots have a higher subscription price?

This whole “they copied it into a gpu” doesn’t matter. People read and interpret. Robots read and interpret. I don’t want to live in a world where every specific device and use needs to be licensed. Especially ex post facto. That will suck so hard.

Re:New York Times considers legal action against OpenAI as c... by version_five on Wednesday August 16, 2023 @07:47PM (#37155392)

  Should I pay more if I memorize the paper? Should I pay more if I read it to my sick friend in the hospital?
It's not even close to either of those things. It's "should I pay more if it infinitesimally affects my perception of english grammar or knowledge of a subject". The llm isn't, for any functional purpose, memorizing, it's getting weight updates as it learns from these examples which are teaspoons of information in a sea of trillions of tokens.
Re:New York Times considers legal action against OpenAI as c... by twoodfin on Wednesday August 16, 2023 @07:48PM (#37155404)
How about “they copied it into a terrific text-to-speech engine and sold ads against the resulting podcasts”?
Re:New York Times considers legal action against OpenAI as c... by prepend on Thursday August 17, 2023 @04:47AM (#37158901)
Reading a text has been ruled as performance and a copyright violation.

I think the issue is that LLMs don’t make a copy or distribute a copy. They use the content to create something else. I don’t remember the copyright term for whether is is transformative enough. But it basically says I can’t copy Starry Night, but I can create a painting with the same color scheme and themes as long as it’s different enough from the original.

Re:New York Times considers legal action against OpenAI as c... by rurp on Wednesday August 16, 2023 @08:04PM (#37155516)
Are you using this knowledge to produce a product that materially undercuts NYT revenue?
Re:New York Times considers legal action against OpenAI as c... by prepend on Thursday August 17, 2023 @04:48AM (#37158910)
That doesn’t matter though. If I licensed material why should it count if I compete or not.

If I watch Lebron James play and use it to develop an athletic training program, does it matter if I play baseball? Or WNBA? Or does it only matter if I play against him in the championship?

Re:New York Times considers legal action against OpenAI as c... by layer8 on Thursday August 17, 2023 @06:55AM (#37159767)
Licenses are bound to the purposes specified by the license. So go read the fine print.
Re:New York Times considers legal action against OpenAI as c... by paulmd on Friday August 18, 2023 @12:06AM (#37171296)
Copyright doesn’t allow you to impose a contract of adhesion to viewers of material you’re broadcasting out into the public. I don’t agree to your contract just because it’s coming into my radio, you can’t create a term of service that imposes additional restrictions on what I can do with it. Copyright may apply but that doesn’t give you the right to enforce additional contracts of adhesion simply based on consuming the content.
Re:New York Times considers legal action against OpenAI as c... by test098 on Wednesday August 16, 2023 @08:45PM (#37155768)
a subscription doesn't give you an automatic escape hatch out of copyright law.

here's their ToS, which is pretty clear about what you cannot do: https://help.nytimes.com/hc/en-us/articles/115014893428-Term... (relevant parts below)

Without NYT’s prior written consent, you shall not:

...

(2) use robots, spiders, scripts, service, software or any manual or automatic device, tool, or process designed to data mine or scrape the Content, data or information from the Services, or otherwise use, access, or collect the Content, data or information from the Services using automated means;

(3) use the Content for the development of any software program, including, but not limited to, training a machine learning or artificial intelligence (AI) system.

...

(5) cache or archive the Content (except for a public search engine’s use of spiders for creating search indices);

Re:New York Times considers legal action against OpenAI as c... by jasomill on Wednesday August 16, 2023 @09:41PM (#37156124)
Full text New York Times articles are available through subscription services other than nytimes.com.

As an example, my local library offers full-text NYT articles through both nytimes.com and ProQuest.

Notably, ProQuest's terms only explicitly ban scraping metadata and developing software or services that "compete or interfere" with ProQuest products:

https://about.proquest.com/en/about/terms-and-conditions

Re:New York Times considers legal action against OpenAI as c... by test098 on Wednesday August 16, 2023 @10:26PM (#37156434)
and your local library and ProQuest are also bound by the same laws, even if they have an existing licensing agreement. from the ToS you just linked (note use of the terms "licensor" and "third party", which would be the NYT in this case):

> Restrictions. Except as expressly permitted above, Customer and its Authorized Users shall not:

> Remove any copyright and other proprietary notices placed upon the Service or any materials retrieved from the Service by ProQuest or its licensors;

> Perform automated searches against ProQuest’s systems (except for non-burdensome federated search services), including automated “bots,” link checkers or other scripts;

> Provide access to or use of the Services by or for the benefit of any unauthorized school, library, organization, or user;

> Publish, broadcast, sell, use or provide access to the Service or any materials retrieved from the Service in any manner that will infringe the copyright or other proprietary rights of ProQuest or its licensors;

> Download all or parts of the Service in a systematic or regular manner or so as to create a collection of materials comprising all or a material subset of the Service, in any form.

> Store any information on the Service that violates applicable law or the rights of any third party.

Re:New York Times considers legal action against OpenAI as c... by paulmd on Friday August 18, 2023 @12:11AM (#37171321)
Grandparent is right that people keep conflating copyright and licensing. That term doesn’t allow you to violate the copyright of a licensor, but, it also doesn’t rule out (eg) fair use.

Fair use would have to be blocked via a license, and it’s going to be difficult to argue that someone agrees to a license merely by turning on their radio. Responding to unauthenticated internet requests with content is the internet equivalent of broadcast and similarly the LinkedIn case held that this did not allow LinkedIn to impose terms of service in a contract of adhesion in this fashion.

Re:New York Times considers legal action against OpenAI as c... by bryanrasmussen on Thursday August 17, 2023 @04:06AM (#37158596)
actually if you had a subscription you might be more screwed than if you crawled it with free access, since having the subscription means agreement to the TOS.
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:18AM (#37158662)
Violating their terms of service doesn't really matter in terms of copyright law though. The statistical properties of that text (which is what the engine snarfs in) aren't protected by copyright.
Re:New York Times considers legal action against OpenAI as c... by ben_w on Thursday August 17, 2023 @04:59AM (#37159003)
> This whole “they copied it into a gpu” doesn’t matter. People read and interpret. Robots read and interpret. I don’t want to live in a world where every specific device and use needs to be licensed. Especially ex post facto. That will suck so hard.

I think you already live in that world (though IANAL), there was a ruling that the Glider cheat tool for WoW was a copyright violation even though it was poking around inside the local copy necessarily made in RAM as part of normal usage of WoW.

https://arstechnica.com/gaming/2009/01/judges-ruling-that-wo...

Re:New York Times considers legal action against OpenAI as c... by kolinko on Thursday August 17, 2023 @03:30AM (#37158375)
Google does the same to produce a search index.
Re:New York Times considers legal action against OpenAI as c... by twelve40 on Thursday August 17, 2023 @04:11AM (#37158620)
you can easily opt out of that, or control it to your heart's desire (including what snippets to show), and it will be honored. There is no way to opt out of this bullshit. So no, not the same at all.
Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:21AM (#37158693)
That's a courtesy of the search engine, not a requirement of copyright.
Re:New York Times considers legal action against OpenAI as c... by dragonwriter on Wednesday August 16, 2023 @07:09PM (#37155095)
Copyright law. Which already covers both mechanical duplication, including mechnaical duplication with automatic alterations to evade detection while continuing to reproduce protect elements if the original.
Re:New York Times considers legal action against OpenAI as c... by thrwayaistartup on Thursday August 17, 2023 @09:16AM (#37161087)
> including mechnaical duplication with automatic alterations to evade detection while continuing to reproduce protect elements if the original.

That's super interesting and is news to me. Thanks for sharing. Would you mind linking to relevant statutes or court decisions?

(This isn't a "citation needed" post -- I believe you, and I'm genuinely curious to read more, but can't find anything!)

Re:New York Times considers legal action against OpenAI as c... by VWWHFSfQ on Wednesday August 16, 2023 @07:32PM (#37155283)
Guidelines for

Assuring

Responsible

Behavior in

AI

Generated

Expressions

G.A.R.B.A.G.E.

Re:New York Times considers legal action against OpenAI as c... by FrustratedMonky on Wednesday August 16, 2023 @07:09PM (#37155092)
If a human reads something, it goes into their brain, and it becomes an influence on future works they produce.

This doesn't mean that 'copywrite' extends into my brain. A company can't copywrite what I'm thinking about. And what if I do try to paraphrase something from memory, from a few sources, and happen to spit out a very similar sentence from memory. Am I breaking the law?

To go further. Since all knowledge is pretty much fed into a human from hundreds of books, movies, TV, internet, all pumped into a human from birth. Then everything in the brain is a product of something with a copywrite. So anything produced is some amalgamation of copywrites.

Why not use similar argument for AI. It is clear when asking it to do something like "write a screen play for Othello using dialog like Tarantino, but with bit of style like Baz Luhrmann". That what it produces is 'as unique as a human' would be, or just as filled with things that have copywrites.

Re:New York Times considers legal action against OpenAI as c... by nullindividual on Wednesday August 16, 2023 @07:13PM (#37155120)
> Am I breaking the law?

The intent of [US] copyright law is to promote new works of art (which can be derivative). So copyright did exactly what it is supposed to do in your analogy. Plus, you're human, which gives you special rights that software doesn't posses.

Re:New York Times considers legal action against OpenAI as c... by FrustratedMonky on Wednesday August 16, 2023 @07:24PM (#37155207)
But. I am allowed to at least read the copywritten material, from which it goes into my brain to become mixed up with everything else, and spit out to produce something 'new' or 'newish'.

Some of these lawsuits are trying to prevent the AI from even 'reading' the material. It can't even be used as an influence.

Wouldn't it be better to treat the products of the AI with the same laws as humans. If the new 'product' is 'too close' to something existing, then they get sued. Just like a musician that has a song with a few notes that sound a little too close to someone's song from 30 years ago gets sued. The songwriter was allowed to listen to the music, it went into their brain and became an influence. If that influence becomes too great, then it can be sued.

Re:New York Times considers legal action against OpenAI as c... by throwaway3618 on Wednesday August 16, 2023 @07:40PM (#37155334)
Seems copyright works in part because of the effort involved in producing something that's similar to something else. You can't just copy and it takes effort to make something different enough, that gives the original a bit of a "moat".

If you take away enough of that effort, the investment in new stuff becomes unviable, perhaps.

Re:New York Times considers legal action against OpenAI as c... by FrustratedMonky on Thursday August 17, 2023 @12:58PM (#37164701)
Agree. I think the 'effort' reduction is the real difference that makes AI products different from human products.

Though this seems different issue from copywrite .

Like, we see that this can impact society, so we need some new laws to guard this.

Re:New York Times considers legal action against OpenAI as c... by lokar on Wednesday August 16, 2023 @07:37PM (#37155318)
Why would that be better? You seems to assume that because these models have some utility to their creators they should be allowed, even when they have negative utility to others who’s work they consume. Why should that be true?
Re:New York Times considers legal action against OpenAI as c... by nullindividual on Thursday August 17, 2023 @08:54AM (#37160831)
> I am allowed to at least read the copywritten material, from which it goes into my brain to become mixed up with everything else, and spit out to produce something 'new' or 'newish'.

Yes, because you're A) human and B) that is how copyright is supposed to work.

AI doesn't enjoy the rights of people. AI is a "talking book" and copying, storing, then repeating someone else's work from your talking book would (likely) run afoul of [US] copyright law.

Re:New York Times considers legal action against OpenAI as c... by alickz on Friday August 18, 2023 @08:08AM (#37174534)
I don't think there's been a tool in the history of mankind with such potential for new works of art than AI tools
Re:New York Times considers legal action against OpenAI as c... by staticman2 on Wednesday August 16, 2023 @07:26PM (#37155225)
"If a human reads something, it goes into their brain"

Humans aren't property. LLM models are. So the comparison is irrelevant and I'll stop you right there.

Re:New York Times considers legal action against OpenAI as c... by BudaDude on Wednesday August 16, 2023 @07:35PM (#37155303)
When you think about it like that, if LLM's are based on the human brain, did we basically reinvent slavery?
Re:New York Times considers legal action against OpenAI as c... by throwaway3618 on Wednesday August 16, 2023 @07:41PM (#37155348)
Not just for the LLMs
Re:New York Times considers legal action against OpenAI as c... by BudaDude on Wednesday August 16, 2023 @07:50PM (#37155424)
True, any neural network. This is a very grey area and rushing to call it "property" ignores what the technology is emulating.
Re:New York Times considers legal action against OpenAI as c... by nonameiguess on Wednesday August 16, 2023 @07:38PM (#37155320)
> And what if I do try to paraphrase something from memory, from a few sources, and happen to spit out a very similar sentence from memory. Am I breaking the law?

If you're doing this for a commercial purpose, yes. Recording artists have been successfully sued for accidentally reusing a melody they claim to not remember ever hearing, provided it really does sound sufficiently similar to the original.

Re:New York Times considers legal action against OpenAI as c... by bubblethink on Wednesday August 16, 2023 @07:55PM (#37155456)
>If, when someone searches online, they are served a paragraph-long answer from an AI tool that refashions reporting from The Times, the need to visit the publisher's website is greatly diminished, said one person involved in the talks.

If, when someone reads a newspaper, they are served a paragraph-long answer from an NYTimes reporter that refashions reporting from local sources, the need to interact with the local sources is greatly diminished.

Re:New York Times considers legal action against OpenAI as c... by belltaco on Wednesday August 16, 2023 @08:15PM (#37155590)
So? The local source is free to sue the NYT for copyright infringement if they so wish.
Re:New York Times considers legal action against OpenAI as c... by ben_w on Thursday August 17, 2023 @04:51AM (#37158928)
This certainly happens in the UK with different newspapers, and starting with a letter demanding payment, but the same idea.

Very few people bother doing this for much the same reason very few bother fighting any of the other terrible decisions made by corporations with legal departments whose annual cost exceeds their personal lifetime earnings.

Re:New York Times considers legal action against OpenAI as c... by Legend2440 on Thursday August 17, 2023 @03:50AM (#37158506)
Not to mention there are already entire news outlets that exist to summarize paywalled journalism. Ever read an article that starts with "The NYTimes reports..."?
Re:New York Times considers legal action against OpenAI as c... by rurp on Wednesday August 16, 2023 @07:59PM (#37155479)
I don't think it's an exaggeration to say that LLMs might lead to the end of the open web, or at least a drastically reduced version of it. So much of these model's utility is in directly competing with the producers of the training data. Content creators and aggregators are seeing more and more reason to restrict and limit access, to avoid having AI companies consume all of their data and then be the ones making money from it going forward.

I fear that LLMs are going to cause the internet to be a much worse and less open space.

Re:New York Times considers legal action against OpenAI as c... by buildbot on Wednesday August 16, 2023 @08:13PM (#37155582)
No it's the death of the corporate content hosting web - the open web was never about making money with your blog post/irc chat/usenet group/etc content, at least in my opinion.

Let data be free! If someone wants to use it to make money, well, it's open, just like open source. It's still not okay to take open source work and claim it as your own, which is what copyright should be limited to. Stealing a photo or plagiarizing an essay is intrinsically different than just having a copy read by something, be it human or an mechanical process such as training a LLM.

Re:New York Times considers legal action against OpenAI as c... by sundaeofshock on Wednesday August 16, 2023 @10:57PM (#37156618)
Maybe I don’t want my non-corporate art to be a part of some large corporation’s training data.
Re:New York Times considers legal action against OpenAI as c... by k_vi on Thursday August 17, 2023 @03:15AM (#37158285)
User-agent: GPTBot Disallow: /
Re:New York Times considers legal action against OpenAI as c... by prmoustache on Thursday August 17, 2023 @03:37AM (#37158430)
ATMO you shouldn't have to maintain knowledge of what kind of crawler bot exist and having to maintain deny list. It should be the opposite, only expressedly allowed content should be crawled by mainaining allow lists.
Re:New York Times considers legal action against OpenAI as c... by wraptile on Thursday August 17, 2023 @04:07AM (#37158599)
You can do the opposite since the inception of robots.txt: User-agent: * Disallow: / and then whitelist google bot and whatnot. Most of the web is already configured this way. Just check robots.txt of any major website, e.g. https://twitter.com/robots.txt
Re:New York Times considers legal action against OpenAI as c... by xnx on Thursday August 17, 2023 @12:34PM (#37164294)
The Allow: directive was an extension to robots.txt added later.
Re:New York Times considers legal action against OpenAI as c... by jamilton on Thursday August 17, 2023 @03:35AM (#37158408)
That was my gut reaction too, but presumably unless it becomes regulated, at least some competitors to OpenAI won't respect any robots.txt and thus any open content might be training data.
Re:New York Times considers legal action against OpenAI as c... by gorbachev on Thursday August 17, 2023 @04:05AM (#37158592)
User-Agent: <new technology category>Bot
Re:New York Times considers legal action against OpenAI as c... by andybak on Thursday August 17, 2023 @04:02AM (#37158574)
Once again I have to point out that LLMs are not and will not be restricted to "large corporations". Even if initial training is prohibitively expensive, open base models already exist.

"corporate this" and "corporate that" are handy rhetorical flourishes but they distort the debate.

Re:New York Times considers legal action against OpenAI as c... by twelve40 on Thursday August 17, 2023 @04:07AM (#37158597)
if you want to give away your data for free, do it, but speak for yourself, not everyone is in position to be able to do it. Just like you can't force other people to give away their software (or anything else, really) for free if they choose to charge for it.
Re:New York Times considers legal action against OpenAI as c... by Riverheart on Thursday August 17, 2023 @09:40AM (#37161390)
[delayed]
Re:New York Times considers legal action against OpenAI as c... by lostmsu on Thursday August 17, 2023 @03:40AM (#37158439)
Let's assume that happens. How do I hedge against it? Is there a convenient way to mirror the bits of the web that are open now? Perhaps a mirror of archive/WayBack machine that could be viewed locally similar to Wikipedia dumps?
Re:New York Times considers legal action against OpenAI as c... by paulmd on Thursday August 17, 2023 @11:38PM (#37171163)
Yes, commoncrawl indexes
Re:New York Times considers legal action against OpenAI as c... by lostmsu on Friday August 18, 2023 @11:08PM (#37185084)
100TB. Dang! I will need to expand my storage significantly to mirror it.
Re:New York Times considers legal action against OpenAI as c... by paulmd on Saturday August 19, 2023 @01:18AM (#37185764)
100TB so far ;)

And that's probably just 1 year. There's data that's vanished over time. The point of all this is that they're point-in-time references of the internet.

the more practical thing is that (the royal) they have set up a rclone like utility that allows you to retrieve a slice of the whole snapshot, if you know your domain.

it would be extremely funni if commoncrawl was a de-facto ipfs, allowing queries along domains (or arbitrary dimensions/clustering) of interest.

Re:New York Times considers legal action against OpenAI as c... by mike_hearn on Thursday August 17, 2023 @03:40AM (#37158443)
Really? Seems like it's not so different conceptually to search engines. They also make money by indexing "training data" of a sort. But the open web trundles on regardless, because the search engines found ways to cut the content producers in on it. I see no reason why that can't be the case here too, with AI companies training their models to act more like search engines when data comes from certain sources - i.e. they'll point you in the right direction but not directly answer you. That will suck for the AI users, but, for many questions the right answer will be available in open or bulk licensable training sets anyway. For example you can get legal access to nearly all books by doing deals with publishers, as Google Books has demonstrated, you can get access to map data by generating it yourself or doing deals with digital map companies etc.
Re:New York Times considers legal action against OpenAI as c... by bryanrasmussen on Thursday August 17, 2023 @04:02AM (#37158577)
>Really? Seems like it's not so different conceptually to search engines. They also make money by indexing "training data" of a sort.

OK, so if a writer X has a blog to put up samples of their work to drive people to buy books and to get writing assignments and someone uses ChatGPT to write something in the style of X - this naively seems like a hit on that author's ability to sell their skills.

And I don't think it is fixable by making the AI act more like a search engine.

Re:New York Times considers legal action against OpenAI as c... by mike_hearn on Thursday August 17, 2023 @04:27AM (#37158747)
Seems like a rerun of the argument over snippets, or the "answer onebox" as Google used to call it where the info you need is directly inlined in the SERP rather than being behind a link.
Re:New York Times considers legal action against OpenAI as c... by bryanrasmussen on Thursday August 17, 2023 @05:10AM (#37159089)
it is exactly a rerun of that argument, except that the UX of AI is different.

Basically AI is structured to make every result an answer onebox whereas in search this only happens sometimes.

Re:New York Times considers legal action against OpenAI as c... by rurp on Thursday August 17, 2023 @11:37AM (#37163326)
Yes I would say that is very similar. Many sites saw their revenue crater, based on data they themselves had created.
Re:New York Times considers legal action against OpenAI as c... by ben_w on Thursday August 17, 2023 @04:45AM (#37158886)
> this naively seems like a hit on that author's ability to sell their skills

If the AI is good, then the author's economic outlook is bad regardless of style.

Regardless of if the AI is or isn't good, then I can still see it being brand damaging, but my gut feeling (IANAL) is that this is more of a trademark issue than a copyright issue, as it's passing off as something it isn't.

Re:New York Times considers legal action against OpenAI as c... by josephg on Thursday August 17, 2023 @04:45AM (#37158891)
It would work both ways. I could also imagine someone asking: “what books should I read to learn about X?” And the LLM could drive sales toward that author’s books.

It’s not clear to me that having their content “unindexed” is good for authors. It’s probably good and bad at the same time.

Re:New York Times considers legal action against OpenAI as c... by twelve40 on Thursday August 17, 2023 @04:04AM (#37158588)
yes indeed, search engines settled that problem a long time ago in a civilized way. The more of these lawsuits, the harder the wild west of "training" will have to search for an acceptable solution to what is currently web-scale theft of data.
Re:New York Times considers legal action against OpenAI as c... by wraptile on Thursday August 17, 2023 @04:00AM (#37158567)
Is this open web in the room with us? Cheekiness aside, the myth of beautiful open web always seems to be exagerated. Most content on the web is already generated, ugly and spammy. Most of the traffic is already owned by mega corporations.

One could easily argue that it's unlikely it'll get worse - if anything, AI could empower competition as now a group of 3 passionate, free writers can compete with agenda-driven, for profit corporations on a similar level. This could very well make the web more open and free as it makes the web more accessible.

Re:New York Times considers legal action against OpenAI as c... by ben_w on Thursday August 17, 2023 @04:24AM (#37158724)
> Most content on the web is already generated, ugly and spammy.

Now? Yes. I grew up when it was just ugly, and by ugly I mean animated gif backgrounds with obvious seams on the tile boundary.

*old man shakes fist at The Cloud*

Re:New York Times considers legal action against OpenAI as c... by bubblethink on Wednesday August 16, 2023 @08:02PM (#37155499)
I think all OpenAI needs to do is scan physical newspapers and OCR them. No ToS to agree to, and no ToS on print editions.
Re:New York Times considers legal action against OpenAI as c... by test098 on Wednesday August 16, 2023 @08:40PM (#37155736)
ignorance of copyright law won't save you here. i can't legally torrent a copywrited music file just because "no tos to agree to" when i listen to it.
Re:New York Times considers legal action against OpenAI as c... by bubblethink on Wednesday August 16, 2023 @08:51PM (#37155794)
There is no copyright involved in training a model. No precedent at least. Only online ToS/API restrictions exist for scraping content. The DMCA issues exist entirely on the (re)distribution side of coyright material. So seeding in a torrent swarm = redistribution. If you download some copyright material somehow and don't share it with anyone, there is no caselaw that says anything about it.
Re:New York Times considers legal action against OpenAI as c... by test098 on Wednesday August 16, 2023 @10:37PM (#37156499)
the NYT certainly seems to think there's a copyright issue - namely, their copyrighted work being directly used to create a potential competitor.

> If you download some copyright material somehow and don't share it with anyone, there is no caselaw that says anything about it.

this is still a violation of the law. you cannot download copyrighted material against the terms of the copyright holder (like downloading a movie or album).

Re:New York Times considers legal action against OpenAI as c... by bandrami on Thursday August 17, 2023 @04:26AM (#37158735)
RIAA kept running into this problem which is why it only sent the lawyers after people they could confirm seeded.
Re:New York Times considers legal action against OpenAI as c... by jiscariot on Wednesday August 16, 2023 @08:03PM (#37155507)
My concern isn't copyright law, but that if trained on the NYT, these llms are going to be favorable to starting conflicts in the middle-east.

https://fair.org/home/20-years-later-nyt-still-cant-face-its...

Re:New York Times considers legal action against OpenAI as c... by exabrial on Thursday August 17, 2023 @08:57PM (#37170302)
Good. Literally anyone’s copyrighted comments on the internet should get a settlement
Re:New York Times considers legal action against OpenAI as c... by CatWChainsaw on Tuesday August 22, 2023 @07:27AM (#37221201)
"I'll keep saying it every time this comes up.

I LOVE being told by techbros that a human painstaking studying one thing at a time, and not memorizing verbatin but rather taking away the core concept, is exactly the same type of "learning" that a model does when it takes in millions of things at once and can spit out copyrighted writing verbatim."

Personally I think they argue that way because they get off on being contrarian out of spite, but to me it's just a signal of maliciousness and stupidity all at once.