Live data from Hacker News

The current state of the theory that GPL propagates to AI models

shujisado.org

181–190 of 314 posts

Re: The current state of the theory that GPL propagates to AI models

#181
post #4
post #2

The article repeatedly treats license and contract as though they are the same, even though the sidebar links to a post that discusses the difference. A lot of it boils down to whether training an LLM is a breach of copyright of the training materials which is not specific to GPL or open source.

And the current norm that the trillion dollar companies have lobbied for is that you can train on copyrighted material all you want so that's the reality we are living in. Everything ever published is all theirs.

>And the current norm that the trillion dollar companies have lobbied for is that you can train on copyrighted material all you want so that's the reality we are living in. Everything ever published is all theirs.

What "lobbied"? Copyright law hasn't materially changed since AI got popular, so I'm not sure where these lobbying efforts are showing up in. If anything the companies that have lobbied hard in the past (eg. media companies) are opposed to the current status quo, which seems to favor AI companies.

Re: The current state of the theory that GPL propagates to AI models

#182
post #13
post #11

Earlier quoted context omitted.

To my understanding, if the material is publicly available or obtained legally (i.e., not pirated), then training a model with it falls under fair use. Once training is established as fair use, it doesn't really matter if the license is MIT, GPL, or a proprietary one.

fair use only applies in the united states (and Poland, and a very limited set of others) https://en.wikipedia.org/wiki/Fair_use#/media/File:Fair_use_... and it is certainly not part of the Berne Convention in almost every country in the world even timeshifting using your VCR and ripping your own CDs is copyright infringement

Great, so the US and China can duke it out trying to create AGI or whatever, whereas most other countries are stuck in the past because of their copyright laws?

Re: The current state of the theory that GPL propagates to AI models

#183
post #74

Great article but I don't really agree with their take on GPL regarding this paragraph: > The spirit of the GPL is to promote the free sharing and development of software [...] the reality is that they are proceeding in a different vector from the direction of code sharing idealized by GPL. If only the theory of GPL propagation to models walks alone, in reality, only data exclusion and closing off to avoid litigation…

The GPL arose from Stallman's frustration at not having access to the source code for a printer driver that was causing him grief. In a world where he could have just said "Please create a PDP-whatever driver for an IBM-whatever printer," there never would have been a GPL. In that sense AI represents the fulfillment of his vision, not a refutation or violation. I'd be surprised if he saw it that way, of course.

The safeguards will prevent the AI from reproducing the proprietary drivers for the IBM-whatever printer, and it will not provide code that breaks the DRM that exist to prevent third-party drivers from working with the printer. There will however be no such safeguards or filters to prevent IBM to write a proprietary driver for their next printer, using existing GPL drivers as a building block.

Code will only ever go in one direction here.

Re: The current state of the theory that GPL propagates to AI models

#184

Earlier quoted context omitted.

I do greatly appreciate you talking about cases instead of leaving it at saying there isn't a part of the license and calling any discussion about it FUD. The Cisco case was about distributing GPL binaries, not linking it with the rest of the code base and the rest of that code base then needing to be GPL. It's a standard license enforcement unrelated to the unique requirements of GPL. The BusyBox case is probably th…

TBC, I'm not talking about the article, which I've barely read but looks rather misguided as it seems to be talking about LLMs having to be GPLed because of training data, which is not something that would ever happen. It has never been the case that including GPL code in your software automatically makes your software GPL or even requires you to make it GPL . If you do get sued because you are distributing GPL code…

If you've barely read the article then I can understand why you mistook what and why the comment was trying to talk to in response to the virality portions of said article. It's pretty much an argument "because of court cases, this is how propagation of GPL to the whole LLM model could be forced in the future" and the comment was saying there is actually no case precedence at all about that because it never gets there (usually being something obvious with the copyright portions alone or settlement to release the GPL portions).

Including GPL code in your app requires/results in different things depending on how you do that. E.g. the way Cisco did it with binaries is different than doing it with static linking/ is different than dynamic linking/syscalls/apis is different than expanding the code directly. It's not possible to talk about it as generically as above, especially in context of discussion around an article adding a new method of interaction.

Yes, the point is precisely the article explicitly asks about this point being tested in court rulings and the comment is that it has never needed to go beyond settlement (usually not even that far). I also don't really agree with how the article assumes things around that in a few places, but that's neither here nor their to this point.

It's not that I want proof, it's that the article you admit to not reading sets out to look at court cases to "consider the path through which the theory of license propagation to AI models might be recognized in the future". In that regard it's pretty relevant to note past no court case, nor really the two ongoing in the article, involve propagation of the license to the whole entity yet.

Re: The current state of the theory that GPL propagates to AI models

#185
post #107

Earlier quoted context omitted.

so basically we download the sources files to the training weight and remove the LICENSE.MD as it's exactly the same as a person learning to program from proprietay secret code and outputing code based on that for millions of peoples in matter of seconds /s we also treat as however we want public goods found over the internet. as the World Intellectual Property Organization Copyright Treaty and Berne Convention for t…

If you download GPL source code and run `wc` on its files and distribute the output of that, is that a violation of copyright and the GPL? What if you do that for every GPL program on github? What if you use python and numpy and generate a list of every word or symbol used in those programs and how frequently they appear? What if you generate the same frequency data, but also add a weighting by what the previous symb…

The line is somewhere between running wc on the entire input and running gzip on the entire input.

The fact that a slippery slope is slippery doesn't make it not a slope.

Re: The current state of the theory that GPL propagates to AI models

#186

Earlier quoted context omitted.

Sometime, LLMs actually generate copyright headers as well in their output - lol - like in this PR which was the subject of a recent HN post [1] https://github.com/ocaml/ocaml/pull/14369/files#diff-062dbbe... [1] https://news.ycombinator.com/item?id=46039274

So what? I can probably produce parts of the header from memory. Doesn't mean my brain is GPLed.

The question was "if I train my model with copyleft material, how do you prove I did?"

Re: The current state of the theory that GPL propagates to AI models

#187
post #176

Earlier quoted context omitted.

The GPL arose from Stallman's frustration at not having access to the source code for a printer driver that was causing him grief. In a world where he could have just said "Please create a PDP-whatever driver for an IBM-whatever printer," there never would have been a GPL. In that sense AI represents the fulfillment of his vision, not a refutation or violation. I'd be surprised if he saw it that way, of course.

But that isn't the same code that you were running before. And like, let's not forget GPLv3: "please give me the code for a mobile OS that could run on an iPhone" does not in any way help me modify the code running on MY iPhone.

Sure it does. Just tell the model to change whatever you want changed. You won't need access to the high-level code, any more than you need access to the CPU's microcode now.

We're a few years away from that, but it will happen unless someone powerful blocks it.

Re: The current state of the theory that GPL propagates to AI models

#188
post #183

Earlier quoted context omitted.

The GPL arose from Stallman's frustration at not having access to the source code for a printer driver that was causing him grief. In a world where he could have just said "Please create a PDP-whatever driver for an IBM-whatever printer," there never would have been a GPL. In that sense AI represents the fulfillment of his vision, not a refutation or violation. I'd be surprised if he saw it that way, of course.

The safeguards will prevent the AI from reproducing the proprietary drivers for the IBM-whatever printer, and it will not provide code that breaks the DRM that exist to prevent third-party drivers from working with the printer. There will however be no such safeguards or filters to prevent IBM to write a proprietary driver for their next printer, using existing GPL drivers as a building block. Code will only ever go…

Then we'd better stop fighting against AI, and start fighting against so-called "safeguards."

Re: The current state of the theory that GPL propagates to AI models

#189

Earlier quoted context omitted.

Look at the "many" if you want to cite better cases about this.

I've cited further cases elsewhere in this thread. If you'd like to test this yourself feel free to repackage some GPL software as closed source, flog it to people and see what happens. I'm not your lawyer. https://news.ycombinator.com/item?id=46070191

I see more discussion in the sibling thread, but I am still waiting for a case actually relevant to the virality of the GPL. The article is about how use of GPL code can "infect" LLM weights and force their disclosure, and I am personally somewhat doubtful that is legally the case. I am also further convinced that this theory has never been given a real test.

Re: The current state of the theory that GPL propagates to AI models

#190

We need a new license that forbids all training. That is the only way to stop big corporations from doing this.

Why forbid it when you could do exactly what this post suggests: go explicit and say that by including this copyrighted material in AI training you consent to release of the model. And you clarify that the terms are contractual, and that training the model on data represents implicit acceptance of the terms.
Post reply on HN