Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Who owns the code Claude Code wrote?

legallayer.substack.com

Post reply on HN
Print view View original

351–360 of 570 posts

Re: Who owns the code Claude Code wrote?

#351

Post by alok-g »

greendestiny wrote: post #314
tensor wrote: post #136

Earlier quoted context omitted.

I still find the idea that "learning" from code is "stealing" kind of ridiculous.

I think that it's absurd that we've jumped to the conclusion backpropagation in neural networks should be legally treated the same as human learning. I mean I don't think think I could find a better description for following the derivatives of error in reproducing a set of works as creating a "derivative work".

>> ... we've jumped to the conclusion backpropagation in neural networks should be legally treated the same as human learning.

I agree. However, the reverse is also likely true, i.e., it cannot currently be denied that learning in humans is different from learning in artificial neural networks from the point of view of production of works that mix ideas/memes from several works processed/read. Surely, as the article says, copyright law talks exclusively about humans, not machines, not animals.

Re: Who owns the code Claude Code wrote?

#352

Post by jongjong »

Arcuru wrote: post #115

Personally, I think that the human directing the agent owns the copyright for whatever is produced, but the ability for the agent to build it in the first place is based off of stolen IP. I'm concerned about the copyright 'washing' this enables though, especially in OSS, and I think the right thing for OSS devs to do is to try to publish resulting code with the strongest copyleft licensing that they are comfortable w…

This interpretation makes sense. I think even the 'fair use' clause in the US doesn't protect LLMs. One argument I've heard often is that LLMs synthesize their training set to produce novel output in the same way as a human would... That may be the case, but legally an LLM isn't a human. You can't look at the output of an LLM and say that it's 'fair use' with respect to its training set; it hasn't been established that AI has the same 'fair use' right as a human does; it's already pushing it that companies have this right (let alone an AI agent); anyway, that's just one problem... Also, this is ignoring the fact that the researchers who compiled the training set COPIED the original copyrighted data in order to produce that training set. They either copied the entire work into the training set or they fed the entire work directly into the LLM; in either case; at some point, the entire work was copied verbatim into the LLM's input layer before it was ingested by the AI. The researchers copied the copyrighted content without permission.

Also, when it comes to code, the case is even more damning because the vast majority of the code which LLMs are trained on was not only copyright but subject to an MIT license (at best) and even the MIT license, which is the most permissive license in existence, still says clearly:

"Permission is hereby granted, free of charge, to any person obtaining a copy of this software"

The word 'person' is used very intentionally here.

I think there should be several kinds of AI taxes which should be distributed to all copyright holders. There should be a tax to go to writers (and book authors), a tax to go to open source developers and a tax for the general population to distribute as UBI to account for small-form content like comments and photography...

People invested a lot of time building their entire careers around the assumption of copyright protection; so for it to be violated on such a scale would be a massive betrayal.

Re: Who owns the code Claude Code wrote?

#353

Post by alok-g »

marcus_holmes wrote: post #349
KallDrexx wrote: post #298

Earlier quoted context omitted.

Do you think that human directing the agent owns copyright for any legal reason? The case Community for Creative Non Violence Vs Reid ( https://en.wikipedia.org/wiki/Community_for_Creative_Non-Vio... ) solidifies a supreme court opinion that someone contracting a work and directing an author does not grant authorship to the commissioner of the work, it grants authorship to the person actually doing the work. The auth…

Interesting, though, that ownership of the code can still be transferred to the employer. So it's in the public domain (because not human authored) but owned by the employer (because the human and/or LLM was employed by the employer)? I don't really understand how this works.

Note: IANAL

I think what this means is that the employee may not be the copyright owner for multiple reasons, which are possibly applicable simultaneously. It does not imply that the employer owns copyright over the work that is in public domain, which would be a contradiction.

Re: Who owns the code Claude Code wrote?

#354

Post by alok-g »

ako wrote: post #191
Arcuru wrote: post #115

Personally, I think that the human directing the agent owns the copyright for whatever is produced, but the ability for the agent to build it in the first place is based off of stolen IP. I'm concerned about the copyright 'washing' this enables though, especially in OSS, and I think the right thing for OSS devs to do is to try to publish resulting code with the strongest copyleft licensing that they are comfortable w…

I've created my own DSL, and instruct Claude Code how to generate code for this DSL using skills. Since this is a new language, and not documented on the web nor on Github, Claude's ability is not based off of stolen IP. At best it's trained on other language concepts, just like we can train ourselves on code on GitHub. Maybe a good reason to create a new programming language?

Interesting, but I still do not think this is as easy. The AI model is still trained on some existing works, and it is generating code in the new DSL or programming language still based some higher level ideas and expressions it has consumed during training. You have added just one more level of indirection. The output cannot anymore be verbatim copy of some existing work or non-short snippets, however, the output may still carry "expression" that are substantially similar to something pre-existing.

Note: IANAL. The above is just from my current understanding.

Re: Who owns the code Claude Code wrote?

#355

Post by alok-g »

jacquesm wrote: post #142
Arcuru wrote: post #115

Personally, I think that the human directing the agent owns the copyright for whatever is produced, but the ability for the agent to build it in the first place is based off of stolen IP. I'm concerned about the copyright 'washing' this enables though, especially in OSS, and I think the right thing for OSS devs to do is to try to publish resulting code with the strongest copyleft licensing that they are comfortable w…

No, that human owns the copyright on the prompt, not on the work product.

If that were true, a developer may own copyright over the source code, but nothing on the compiled binaries, and I could download practically all software available as compiled binaries and use for free.

Re: Who owns the code Claude Code wrote?

#356

Post by alok-g »

varispeed wrote: post #132
Arcuru wrote: post #115

Personally, I think that the human directing the agent owns the copyright for whatever is produced, but the ability for the agent to build it in the first place is based off of stolen IP. I'm concerned about the copyright 'washing' this enables though, especially in OSS, and I think the right thing for OSS devs to do is to try to publish resulting code with the strongest copyleft licensing that they are comfortable w…

I find idea that the code could be copyrightable as weak. There are only so many ways to write a for loop. Similarly you can't copyright schematics (apart from exact visual representation as form of art). Code is just a schematic.

Note: IANAL

Copyrights already preclude short phrases for the same reason -- there are only so many ways in which short phrases could be produced. The moment a work becomes larger (large enough; AFAIK, the threshold is not precisely defined), the reasoning you applied fails to apply.

The Google-Oracle lawsuit did not decide whether APIs (when large in number) are copyrightable or not.

Re: Who owns the code Claude Code wrote?

#357

Post by greendestiny »

alok-g wrote: post #351
greendestiny wrote: post #314

Earlier quoted context omitted.

I think that it's absurd that we've jumped to the conclusion backpropagation in neural networks should be legally treated the same as human learning. I mean I don't think think I could find a better description for following the derivatives of error in reproducing a set of works as creating a "derivative work".

>> ... we've jumped to the conclusion backpropagation in neural networks should be legally treated the same as human learning. I agree. However, the reverse is also likely true, i.e., it cannot currently be denied that learning in humans is different from learning in artificial neural networks from the point of view of production of works that mix ideas/memes from several works processed/read. Surely, as the article…

I understand the article - the point about 'learning' is that if the model and its outputs are a derivative works then the copyright belongs to the human creators of the works it was trained on.

Edit*: Or perhaps put more pseudo legally that the created works infringe on the copyrights of the original human creators.

Re: Who owns the code Claude Code wrote?

#359

Post by alok-g »

greendestiny wrote: post #357
alok-g wrote: post #351

Earlier quoted context omitted.

>> ... we've jumped to the conclusion backpropagation in neural networks should be legally treated the same as human learning. I agree. However, the reverse is also likely true, i.e., it cannot currently be denied that learning in humans is different from learning in artificial neural networks from the point of view of production of works that mix ideas/memes from several works processed/read. Surely, as the article…

I understand the article - the point about 'learning' is that if the model and its outputs are a derivative works then the copyright belongs to the human creators of the works it was trained on. Edit*: Or perhaps put more pseudo legally that the created works infringe on the copyrights of the original human creators.

The part I agree to is that copyright law calls out humans specifically as the potential owners of copyright. So what you suggest seems to be the only possibility out. Calling out humans could imply that when a human reads a thousand books and then writes something basis the same but which is not a substantial copy of anything explicitly read, that human owns the copyright to the text written. Whereas, if an artificial neural network does the same (hypothetically writing the same text), it would not.

The above does not follow from, imply or conclude anything about learning in artificial neural networks and humans being similar or dissimilar.

Re: Who owns the code Claude Code wrote?

#360

Post by zarzavat »

KallDrexx wrote: post #298
Arcuru wrote: post #115

Personally, I think that the human directing the agent owns the copyright for whatever is produced, but the ability for the agent to build it in the first place is based off of stolen IP. I'm concerned about the copyright 'washing' this enables though, especially in OSS, and I think the right thing for OSS devs to do is to try to publish resulting code with the strongest copyleft licensing that they are comfortable w…

Do you think that human directing the agent owns copyright for any legal reason? The case Community for Creative Non Violence Vs Reid ( https://en.wikipedia.org/wiki/Community_for_Creative_Non-Vio... ) solidifies a supreme court opinion that someone contracting a work and directing an author does not grant authorship to the commissioner of the work, it grants authorship to the person actually doing the work. The auth…

It depends on what level of creative control you had over the code.

Code is protected by copyright as a literary work. The method is not protected by copyright, that would be the domain of patents. What's protected are the words.

If you say "Claude, build me a website about X" then you do not have any creative control over the literary work Claude is producing. You just told a machine to write it for you. Nor, like a compiler, is it derivative of any other work that you wrote.

If, on the other hand, you are working jointly with Claude to make specific changes to the code on a line-by-line basis, then you will have no problem claiming copyright over the code. Claude in this case is acting as a tool, but there's still a human making decisions about the code.

In the case where you wrote a bunch of markdown and then told Claude to generate the corresponding code but didn't have any involvement in writing the code itself, you could perhaps claim that the code is a derivative work of the markdown, a court would have to handle that case-by-case basis and evaluate how much control you exerted over the work.

Post reply on HN