Dear all, I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC. However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect. GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations. It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art: https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years. Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house. In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws. There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it. I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on. Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dependent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble. The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory. I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters. I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists. With best regards, Damian
Here is Oracle / OpenJDK's stance on the copyright aspect: https://openjdk.org/legal/ai
The Oracle Contributor Agreement (OCA) requires that a contributor own the intellectual property rights in each contribution and be able to grant those rights to Oracle, without restriction. Most generative AI tools, however, are trained on copyrighted and licensed content, and their output can include content that infringes those copyrights and licenses, so contributing such content would violate the OCA. Whether a user of a generative AI tool has IP rights in content generated by the tool is the subject of active litigation.
Now, I immensely dislike Oracle because they are a lawsuit mill before being a software company, but one has to admit that they have some experience in understanding copyright law (and winning their lawsuits, frivolous as they are). Le 31/07/2026 à 10:09, damiancoding--- via ghc-devs a écrit :
Dear all,
I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC.
However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect.
GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations.
It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art: https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright
Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years.
Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house.
In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws.
There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it.
I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on.
Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dep endent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble.
The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory.
I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters.
I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists.
With best regards, Damian _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
-- Hécate ✨ WWW: https://glitchbra.in RUN: BSD
It did come up in the discussion briefly. I think there are two big aspects to this for GHC: Copyright or lack thereof and Licensing. Lack of Copyright: GHC already uses a permissive license. Loss of copyright on parts of the code base would allow those parts to be used without carrying the BSD3 license forward. But in practice this has little effect. The user would not be able to license it as it's not under copyright at that point anyway. So use would remain permissive, although no longer under BSD3 license. To put this into other words. GHC is already not restricting use of the source. And should LLM generated parts of GHC not be covered by copyright this would not meaningfully change. This concern is primarily relevant for projects that want to *restrict* specific use cases. As the GCC project recently did in order to safeguard the applicability of the (L)GPL license. Licensing/Copyright being held by other parties: Contributions to GHC are expected to be licensed under BSD3 or compatible license. While agree that this is relevant, ultimately the responsibility of only contributing source that is compatible with GHCs licensing falls on the user, independent of the tool used in question. If there is a strong argument that companies could claim copyright over code generated by their models it would indeed be right to reject those. However as you alluded to all current providers are very explicit about granting the user all legal rights and disclaiming the rights for themselves. For example here is the relevant blurp for claude: /As between the parties and to the extent permitted by applicable law, Anthropic agrees that Customer (a) retains all rights to its Inputs, and (b) owns its Outputs. Anthropic disclaims any rights it receives to the Customer Content under these Terms. Subject to Customer’s compliance with these Terms, Anthropic hereby assigns to Customer its right, title and interest (if any) in and to Outputs./ A scenario where those providers would be able to restrict distribution of generated source in a GHC compatible way *after the fact* seems impossible to me given these terms of use. There being one caveat. This only applies if users don't violate the terms of use. So a user (re)generating swaths of code that get embedded into GHC, and then goes on to violate the terms of use could change this. Perhaps that is something worth thinking about? Overall I think the legal risks for GHC are constrained to my caveat above, at least if we ignore laws being arbitrarily rewritten and reinterpreted based on economic incentives. There is little that can be done about the later in terms of policy to protect ourselves so I don't think it should inform policy. For the former perhaps the HF would be willing to sponsor a legal consultation to clarify the possible impact of: * Soundness of the waivers to rights over generated code by the common providers. * A contributor generating code and violating the LLM providers terms of use as above. * The consequences of a possible lack of legal copyright for part of the source code. * Generated code making use liable for copyright violations against parties whos code was part of the training set. I will reach out and see if we can make this happen. Andreas On 31/07/2026 10:09, damiancoding--- via ghc-devs wrote:
Dear all,
I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC.
However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect.
GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations.
It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art:https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright
Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years.
Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house.
In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws.
There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it.
I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on.
Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dep endent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble.
The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory.
I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters.
I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists.
With best regards, Damian _______________________________________________ ghc-devs mailing list --ghc-devs@haskell.org To unsubscribe send an email toghc-devs-leave@haskell.org
On Friday, 31 July 2026, 08:09, Damian wrote:
GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations.
That’s not really the situation of a person without citizenship, is it? In Estonia, where I live, there are quite a few such stateless persons: former Soviet Union citizens who acquired neither the Russian nor the Estonian citizenship. I would be very surprised to learn that the Estonian state doesn’t enforce laws and regulations for their benefit. All the best, Wolfgang
Dear Wolfgang, You might be surprised to learn that indeed unrecognized or missing proof of citizenship can be devastating, as for example in this commonly cited example: https://en.wikipedia.org/wiki/Mehran_Karimi_Nasseri This person was effectively without citizenship as they were unable to prove it due to missing documents. This illustrates what I was talking about previously. Legal is full of surprises like these and counter-intuitive twists. You can't use common-sense logic, say "I would be very surprised to learn" and expect that to have much meaning in a legal setting. Probably no one on this list is equipped to make strong predictions about legal outcomes, especially in a changing legal landscape. I commend the first reply for sourcing from a place that might provide some insight. Subsequent to that, all I can say, if we're software developers, please let's stick to software development and not pretend we're lawyers. I would hope the conversation going forward would stay clear of legal opinions unless they come from someone actively practicing law in copyright either in one of the big markets or international copyright, and that person can prove their active credentials. It would be much more constructive to talk about how to find a group of experts like this and how to fund their work. What I can bring to the table is my experience in choosing, directing, and working with legal counsel and knowing how to tell when dealing with good, or bad, counsel, which I have done multiple times in the past. Best regards, Damian
Am Fr 31.07.2026 14:25 schrieb damiancoding--- via ghc-devs:
Dear Wolfgang,
You might be surprised to learn that indeed unrecognized or missing proof of citizenship can be devastating, as for example in this commonly cited example:
https://en.wikipedia.org/wiki/Mehran_Karimi_Nasseri
This person was effectively without citizenship as they were unable to prove it due to missing documents.
That’s a quite specific situation. However, you had made a claim about stateless persons *in general*, saying that “no one will advocate for” them and “no government will back up legitimate claims around” them “by enforcing laws and regulations”. It’s this general claim that I disputed.
Legal is full of surprises like these and counter-intuitive twists. You can’t use common-sense logic, say “I would be very surprised to learn” and expect that to have much meaning in a legal setting.
It’s not just about my intuition here. I was talking about post-Soviet stateless persons in Estonia, of which there are tens of thousands. If they would effectively be without rights, I would have surely heard about that during the 15 years I’ve been living in Estonia. See https://en.wikipedia.org/wiki/Estonian_alien%27s_passport for some information regarding these people.
I would hope the conversation going forward would stay clear of legal opinions unless they come from someone actively practicing law in copyright either in one of the big markets or international copyright, and that person can prove their active credentials.
I usually don’t speak about legal issues, precisely because law is not my field of expertise. However, you had made an apparently wrong claim and had partly based your argumentation on it. Therefore, I considered it important to dispute this claim. Such corrective statements shouldn’t be hindered based on the fact that the persons making them aren’t lawyers. All the best, Wolfgang
Hi Damian, As an asbestos connoisseur (srsly?) and not a scholar of the law, even less across multitudes of jurisdictions. I'd like to posit the following: We have roughly three categories: 1. Restrictive (various form of licenses, EULA, ...) licensing which tries to tell their users what they can, and can not do with the software. 2. Permissive (BSD, MIT, ...) which in spirit say: "Hey here's some code I'm proud of, please mention that your work was derived from this." 3. Copyleft (FSF/Gnu licenses), which kinda say: "Hey, here's some code I've created, and I'm happy for you to use it and modify it, but if you distribute, I'm putting obligations on you". Let's assume there is some violation to any of these: 1. There is probably some body who derives significant financial or other incentives from the licensing, and as such has a strong interest in ensuring said licenses are upheld if violated. 2. This is generally the "throwing code over the fence" situation, take it or leave. If it is violated (e.g. someone didn't add the license/mention) is the author going to sue in court? The whole thing centers around an honor/honest system. Also oftentimes there are no financial incentives attached in a way that would make suing someone for said violation makes much sense. 3. This is a bit of an interesting case. There is a body (FSF) behind these, whose purpose it is to further software freedom, and the Gnu licenses are the vehicle for this. It's also the only OSS license that has seen some court cases. And there are some cases in the West from which one might be able to derive something. Now let's assume someone contributes significant LLM generated code to these, with the assumption that this code is not copyrightable, or potentially falls under copyright violations. 1. This is a problem, as your basis for claiming copyright and subsequently enforcing licenses over code that you potentially can't claim ownership over is tricky, I guess. Legal scholars will certainly have their fun with this, with probably different outcomes across different jurisdictions. 2. Again this license mostly says: hey here is some code I threw over the wall, please include the license and attribute that I curated this code, e.g. provide some lineage. 3. For the copyleft license to work, a much stronger ownership claim needs to be made, e.g. if it is upheld that LLM output is uncopyrightable, and significant parts of the project end up with LLM generated content, the copyright, and thus validity of the copyleft license might be questionable. (How can you enforce something over something you don't own?) This also explains why we see the stratification of LLM policies based on project philosophies, and licenses. 1. A proprietary licenses project with restrictive licenses runs the risk of inadvertently potentially including regurgitated copyleft code; which should courts at some point decide that said code is actually copyleft, might have poisoned the codebase enough to force it under a compatible copyleft license. 2. Somewhat similar problem of potentially poisoning the project to potentially having to adapt a copyleft license. This is really only an issue if there are downstream proprietary consumers. 3. These projects really have an incentive of making sure to stay pure to be able to uphold their copyleft claims. As we see businesses increasingly adopt LLMs in their day-to-day use--whether or not we think this is a good idea--they have a clear incentive to lobby for laws that do not poison their codebases and invalidate their licenses. Similarly for the LLM providers have similar incentives to make sure that the product they sell is compatible with the consumers they are trying to sell their product to. Given that lawyers and lobbying is a rather expensive endeavor, I strongly suspect that we'll see judgements similar to the non-copyright for purely LLM generated output, and human authorship going to play a central role in the US. This of course is just my prediction, not legal advice, given my experience as an asbestos connoisseur. This also seems to be reflecting in the LLM policies we see emerging: 1. corporations seem to make the judgement call that values LLM use above legal risks. And I think they generally have their own legal counsel involved in these decisions. In many corporate settings the use of LLMs seems to be permissible at this point. Restrictions on LLM providers may exist, and some providers offer indemnification in their enterprise offerings. 2. Permissive projects seem to be more relaxed in their LLM policies and mostly focus on the project's culture without being overly prescriptive. Generally seemingly leaning towards human authorship. 3. copyleft projects feeling the need to protect their purity to continue upholding their license claims going forward. There are of course projects in group 2, and group 1 that also opt for less permissive LLM policies for various project reasons. I believe however, on the whole, we do see more restrictive policies from copyleft projects compared to more permissive licenses projects, reflecting a somewhat different understanding of the project's identity around its code. A rather extreme take and thought experiment then is: shouldn't copyleft licensed projects campaign for excessive LLM use in permissive licensed projects, while being very strict about LLM use in their own codebase? This might later open up the claim that substantial portions of copyleft code ended up in other projects, and as such those would have to be relicensed under copyleft licenses, if these claims hold up in court. Therefore if the idealized goal is to make every software free software, not using LLMs oneselves but promoting their use, while hoping for favourable legal outcomes seems a viable although risky strategy. I think the counterpoint to this applies to a permissively licensed project like GHC (as Andreas also outlined), the risk is mostly getting infected by copyleft code enough to force a relicensing under copyleft terms. Ultimately I think the whole licensing topic is very interesting, yet seems to be a different dimension to the questions concerning guidelines and acceptable behaviour around contributions, the initial policy tried to address. As far as I see we have legal, behavioural and ethical/philosophical/ideological concerns around the use of LLMs. I think we would probably do well to separate them out to address them, rather than lumping them together, which is where misunderstandings happen. I don't think there is a natural correlation between these dimensions. I also believe each of these dimensions should be taken seriously. What the conclusion will be, I don't know; it could be that an additional policy, or mission/identity statement, or license or ... will be the end result. The more precise and clear our policies are, the better it will be for the project. Solving something indirectly or by proxy will only lead to uncertainty, confusion, relitigation and debates. Best, Moritz On Fri, 31 Jul 2026 at 15:09, damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Dear all,
I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC.
However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect.
GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations.
It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art: https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright
Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years.
Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house.
In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws.
There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it.
I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on.
Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dep endent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble.
The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory.
I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters.
I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists.
With best regards, Damian _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Moritz Angermann wrote:
Hi Damian,
As an asbestos connoisseur (srsly?) and not a scholar of the law, even less across multitudes of jurisdictions. I'd like to posit the following: We have roughly three categories:
1. Restrictive (various form of licenses, EULA, ...) licensing which tries to tell their users what they can, and can not do with the software. 2. Permissive (BSD, MIT, ...) which in spirit say: "Hey here's some code I'm proud of, please mention that your work was derived from this." 3. Copyleft (FSF/Gnu licenses), which kinda say: "Hey, here's some code I've created, and I'm happy for you to use it and modify it, but if you distribute, I'm putting obligations on you".
Let's assume there is some violation to any of these:
1. There is probably some body who derives significant financial or other incentives from the licensing, and as such has a strong interest in ensuring said licenses are upheld if violated. 2. This is generally the "throwing code over the fence" situation, take it or leave. If it is violated (e.g. someone didn't add the license/mention) is the author going to sue in court? The whole thing centers around an honor/honest system. Also oftentimes there are no financial incentives attached in a way that would make suing someone for said violation makes much sense. 3. This is a bit of an interesting case. There is a body (FSF) behind these, whose purpose it is to further software freedom, and the Gnu licenses are the vehicle for this. It's also the only OSS license that has seen some court cases. And there are some cases in the West from which one might be able to derive something.
Now let's assume someone contributes significant LLM generated code to these, with the assumption that this code is not copyrightable, or potentially falls under copyright violations.
1. This is a problem, as your basis for claiming copyright and subsequently enforcing licenses over code that you potentially can't claim ownership over is tricky, I guess. Legal scholars will certainly have their fun with this, with probably different outcomes across different jurisdictions. 2. Again this license mostly says: hey here is some code I threw over the wall, please include the license and attribute that I curated this code, e.g. provide some lineage. 3. For the copyleft license to work, a much stronger ownership claim needs to be made, e.g. if it is upheld that LLM output is uncopyrightable, and significant parts of the project end up with LLM generated content, the copyright, and thus validity of the copyleft license might be questionable. (How can you enforce something over something you don't own?)
This also explains why we see the stratification of LLM policies based on project philosophies, and licenses.
1. A proprietary licenses project with restrictive licenses runs the risk of inadvertently potentially including regurgitated copyleft code; which should courts at some point decide that said code is actually copyleft, might have poisoned the codebase enough to force it under a compatible copyleft license. 2. Somewhat similar problem of potentially poisoning the project to potentially having to adapt a copyleft license. This is really only an issue if there are downstream proprietary consumers. 3. These projects really have an incentive of making sure to stay pure to be able to uphold their copyleft claims.
As we see businesses increasingly adopt LLMs in their day-to-day use--whether or not we think this is a good idea--they have a clear incentive to lobby for laws that do not poison their codebases and invalidate their licenses. Similarly for the LLM providers have similar incentives to make sure that the product they sell is compatible with the consumers they are trying to sell their product to. Given that lawyers and lobbying is a rather expensive endeavor, I strongly suspect that we'll see judgements similar to the non-copyright for purely LLM generated output, and human authorship going to play a central role in the US. This of course is just my prediction, not legal advice, given my experience as an asbestos connoisseur.
This also seems to be reflecting in the LLM policies we see emerging:
1. corporations seem to make the judgement call that values LLM use above legal risks. And I think they generally have their own legal counsel involved in these decisions. In many corporate settings the use of LLMs seems to be permissible at this point. Restrictions on LLM providers may exist, and some providers offer indemnification in their enterprise offerings. 2. Permissive projects seem to be more relaxed in their LLM policies and mostly focus on the project's culture without being overly prescriptive. Generally seemingly leaning towards human authorship. 3. copyleft projects feeling the need to protect their purity to continue upholding their license claims going forward.
Before we go too far in that direction, what's your Bar ID? Alternatively, if you don't believe in qualifications, can you give me some tips on cardiology, process chemistry, or Finnish grammar?
There are of course projects in group 2, and group 1 that also opt for less permissive LLM policies for various project reasons. I believe however, on the whole, we do see more restrictive policies from copyleft projects compared to more permissive licenses projects, reflecting a somewhat different understanding of the project's identity around its code.
A rather extreme take and thought experiment then is: shouldn't copyleft licensed projects campaign for excessive LLM use in permissive licensed projects, while being very strict about LLM use in their own codebase? This might later open up the claim that substantial portions of copyleft code ended up in other projects, and as such those would have to be relicensed under copyleft licenses, if these claims hold up in court. Therefore if the idealized goal is to make every software free software, not using LLMs oneselves but promoting their use, while hoping for favourable legal outcomes seems a viable although risky strategy.
This isn't relevant to GHC
I think the counterpoint to this applies to a permissively licensed project like GHC (as Andreas also outlined), the risk is mostly getting infected by copyleft code enough to force a relicensing under copyleft terms.
Ultimately I think the whole licensing topic is very interesting, yet seems to be a different dimension to the questions concerning guidelines and acceptable behaviour around contributions, the initial policy tried to address.
Yes, that is my point - there should be discussion of licensing and copyright. The discussion should be had by people who are equipped to do so. I am merely advocating for the discussion to be held.
As far as I see we have legal, behavioural and ethical/philosophical/ideological concerns around the use of LLMs. I think we would probably do well to separate them out to address them, rather than lumping them together, which is where misunderstandings happen. I don't think there is a natural correlation between these dimensions. I also believe each of these dimensions should be taken seriously. What the conclusion will be, I don't know; it could be that an additional policy, or mission/identity statement, or license or ... will be the end result. The more precise and clear our policies are, the better it will be for the project. Solving something indirectly or by proxy will only lead to uncertainty, confusion, relitigation and debates.
Best, Moritz
On Fri, 31 Jul 2026 at 15:09, damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Dear all, I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC. However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect. GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations. It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art: https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years. Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house. In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws. There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it. I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on. Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dep endent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble. The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory. I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters. I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists. With best regards, Damian _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Am Mo 03.08.2026 11:11 schrieb Damian:
Moritz Angermann wrote:
This also seems to be reflecting in the LLM policies we see emerging:
1. corporations seem to make the judgement call that values LLM use above legal risks. And I think they generally have their own legal counsel involved in these decisions. In many corporate settings the use of LLMs seems to be permissible at this point. Restrictions on LLM providers may exist, and some providers offer indemnification in their enterprise offerings.
2. Permissive projects seem to be more relaxed in their LLM policies and mostly focus on the project's culture without being overly prescriptive. Generally seemingly leaning towards human authorship.
3. copyleft projects feeling the need to protect their purity to continue upholding their license claims going forward.
Before we go too far in that direction, what's your Bar ID?
Come on. Moritz was merely writing about typical approaches to LLM use in different settings, speculating about the reasons for choosing these approaches. He was not even making any hard claims about whether these reasons would be valid. Generally, I’d welcome if you wouldn’t tell people to not write certain things based on them not being legal experts, in particular since you’ve already made a wrong statement on a legal matter in this discussion, while at the same time stressing your experience with legal issues. All the best, Wolfgang
Come on.
Just trying to preclude excessive armchair lawyering. The discussion we should be having is how to find the right lawyers and how to fund their work.
Generally, I’d welcome if you wouldn’t tell people to not write certain things based on them not being legal experts
Why not? There's a lot of bad advice and unfounded opinions to be had from non-experts. Programmers especially are well known for believing they are experts at everything, and then they end up re-inventing juicers, busses, trains, etc. It's a well known trend and needs special care.
you’ve already made a wrong statement on a legal matter
If I made a mistake then that of course supports the notion that people will make mistakes, so everything's logically sound. So, to get this back on track: how do we locate the right lawyers? Can the Haskell Foundation even fund them? Does this require a fundraiser of some sort? Could the budgeting and/or fundraiser be joined by other communities and projects with the same issue? LLMs vs copyright is not a uniquely GHC thing. Maybe it's time to reach out to other projects under similar licenses? Other programming languages? Python, Rust, Postgres, PHP, Ruby, etc? Apache foundation? Most of the other large programming communities are corporate-owned, so this is our initial list of options.
I don't know where I have questioned the qualifications. Maybe you can help me point that out! Maybe it's just that people who do not have a qualifying degree in law should have an opinion on such matters relating to the law in the broadest sense? The implied asbestos connoisseur qualification, I will though reject, as I had no formal training in that either (same as law, as I pointed out before). What I do disagree with is that all of this lives in some parallel universe, that is exclusively understandable by people who studied law at length. I know that notion gets pushed every now and then. Maybe I can just disclaim any liability for any contract I ever signed, simply on the basis that, well, it's legal text, and I can't possibly understand what it means as I don't have the necessary degree in law to fully grasp the implied ramifications. Discussing this topic is probably worthwhile. Having experts in the field help out here probably too. What are the reasonable practical options we do have? Can we invoke legal counsel for each and every Merge Request, for all applicable jurisdictions? Which are those? Do we need to reject contributions from people who are living in countries sanctioned by the United States? How would we even reasonably enforce this? Do we think that requesting indemnification from contributors for their contributions is practicable? How does this relate to past contributions? So yes, if someone feels the need and drive to have this discussion with legal scholars of course, having it would certainly be interesting and hopefully beneficial. I guess for legal coverage I should explicitly state in each and every of my emails that: - This is no legal advice! - This is no financial advice! - This email states my personal opinion. - This email does not reflect my employers (if any) opinion. - I do not claim to hold any relevant degrees. and maybe a litany of other disclaimers? On Mon, 3 Aug 2026 at 18:11, damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Moritz Angermann wrote:
Hi Damian,
As an asbestos connoisseur (srsly?) and not a scholar of the law, even less across multitudes of jurisdictions. I'd like to posit the following: We have roughly three categories:
1. Restrictive (various form of licenses, EULA, ...) licensing which tries to tell their users what they can, and can not do with the software. 2. Permissive (BSD, MIT, ...) which in spirit say: "Hey here's some code I'm proud of, please mention that your work was derived from this." 3. Copyleft (FSF/Gnu licenses), which kinda say: "Hey, here's some code I've created, and I'm happy for you to use it and modify it, but if you distribute, I'm putting obligations on you".
Let's assume there is some violation to any of these:
1. There is probably some body who derives significant financial or other incentives from the licensing, and as such has a strong interest in ensuring said licenses are upheld if violated. 2. This is generally the "throwing code over the fence" situation, take it or leave. If it is violated (e.g. someone didn't add the license/mention) is the author going to sue in court? The whole thing centers around an honor/honest system. Also oftentimes there are no financial incentives attached in a way that would make suing someone for said violation makes much sense. 3. This is a bit of an interesting case. There is a body (FSF) behind these, whose purpose it is to further software freedom, and the Gnu licenses are the vehicle for this. It's also the only OSS license that has seen some court cases. And there are some cases in the West from which one might be able to derive something.
Now let's assume someone contributes significant LLM generated code to these, with the assumption that this code is not copyrightable, or potentially falls under copyright violations.
1. This is a problem, as your basis for claiming copyright and subsequently enforcing licenses over code that you potentially can't claim ownership over is tricky, I guess. Legal scholars will certainly have their fun with this, with probably different outcomes across different jurisdictions. 2. Again this license mostly says: hey here is some code I threw over the wall, please include the license and attribute that I curated this code, e.g. provide some lineage. 3. For the copyleft license to work, a much stronger ownership claim needs to be made, e.g. if it is upheld that LLM output is uncopyrightable, and significant parts of the project end up with LLM generated content, the copyright, and thus validity of the copyleft license might be questionable. (How can you enforce something over something you don't own?)
This also explains why we see the stratification of LLM policies based on project philosophies, and licenses.
1. A proprietary licenses project with restrictive licenses runs the risk of inadvertently potentially including regurgitated copyleft code; which should courts at some point decide that said code is actually copyleft, might have poisoned the codebase enough to force it under a compatible copyleft license. 2. Somewhat similar problem of potentially poisoning the project to potentially having to adapt a copyleft license. This is really only an issue if there are downstream proprietary consumers. 3. These projects really have an incentive of making sure to stay pure to be able to uphold their copyleft claims.
As we see businesses increasingly adopt LLMs in their day-to-day use--whether or not we think this is a good idea--they have a clear incentive to lobby for laws that do not poison their codebases and invalidate their licenses. Similarly for the LLM providers have similar incentives to make sure that the product they sell is compatible with the consumers they are trying to sell their product to. Given that lawyers and lobbying is a rather expensive endeavor, I strongly suspect that we'll see judgements similar to the non-copyright for purely LLM generated output, and human authorship going to play a central role in the US. This of course is just my prediction, not legal advice, given my experience as an asbestos connoisseur.
This also seems to be reflecting in the LLM policies we see emerging:
1. corporations seem to make the judgement call that values LLM use above legal risks. And I think they generally have their own legal counsel involved in these decisions. In many corporate settings the use of LLMs seems to be permissible at this point. Restrictions on LLM providers may exist, and some providers offer indemnification in their enterprise offerings. 2. Permissive projects seem to be more relaxed in their LLM policies and mostly focus on the project's culture without being overly prescriptive. Generally seemingly leaning towards human authorship. 3. copyleft projects feeling the need to protect their purity to continue upholding their license claims going forward.
Before we go too far in that direction, what's your Bar ID?
Alternatively, if you don't believe in qualifications, can you give me some tips on cardiology, process chemistry, or Finnish grammar?
There are of course projects in group 2, and group 1 that also opt for less permissive LLM policies for various project reasons. I believe however, on the whole, we do see more restrictive policies from copyleft projects compared to more permissive licenses projects, reflecting a somewhat different understanding of the project's identity around its code.
A rather extreme take and thought experiment then is: shouldn't copyleft licensed projects campaign for excessive LLM use in permissive licensed projects, while being very strict about LLM use in their own codebase? This might later open up the claim that substantial portions of copyleft code ended up in other projects, and as such those would have to be relicensed under copyleft licenses, if these claims hold up in court. Therefore if the idealized goal is to make every software free software, not using LLMs oneselves but promoting their use, while hoping for favourable legal outcomes seems a viable although risky strategy.
This isn't relevant to GHC
I think the counterpoint to this applies to a permissively licensed project like GHC (as Andreas also outlined), the risk is mostly getting infected by copyleft code enough to force a relicensing under copyleft terms.
Ultimately I think the whole licensing topic is very interesting, yet seems to be a different dimension to the questions concerning guidelines and acceptable behaviour around contributions, the initial policy tried to address.
Yes, that is my point - there should be discussion of licensing and copyright. The discussion should be had by people who are equipped to do so. I am merely advocating for the discussion to be held.
As far as I see we have legal, behavioural and ethical/philosophical/ideological concerns around the use of LLMs. I think we would probably do well to separate them out to address them, rather than lumping them together, which is where misunderstandings happen. I don't think there is a natural correlation between these dimensions. I also believe each of these dimensions should be taken seriously. What the conclusion will be, I don't know; it could be that an additional policy, or mission/identity statement, or license or ... will be the end result. The more precise and clear our policies are, the better it will be for the project. Solving something indirectly or by proxy will only lead to uncertainty, confusion, relitigation and debates.
Best, Moritz
On Fri, 31 Jul 2026 at 15:09, damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Dear all, I have thus far been perusing the LLM policy discussion from the bleachers. A lot of conversation has been had about the ethical and practical reasons to support or abolish LLM use for GHC. However, one really important point seems to have been all but omitted from the conversation, and that is the legal aspect. GHC, like any other project using any licensing at all, enjoys protections based in the legal system of every country it is being used and developed in. That is no small thing and is essentially what allows GHC to exist in the legal sense. A body of work untethered from this is like a person without citizenship - no one will advocate for it, and no government will back up legitimate claims around it by enforcing laws and regulations. It is the contention of some legal scholars that having a significant (not large or overwhelming, merely significant) amount of AI based contributions in a body of work can lead to lack of recognition of its copyright in court. This opinion is already backed up by case law relating to art: https://www.theverge.com/policy/887678/supreme-court-ai-art-copyright Such a status would put the project's protections in danger legally and could turn to foil the efforts of GHC contributors over the past numerous years. Of course, art is not code, and I would say many people's code is not art... but absence of proof is not proof of absence. This is merely legal limbo, not legal certainty, and early indications suggest the tree may fall on the house. In fact, who is to say that at least in some jurisdictions, AI companies couldn't go rogue and turn around and say anything that's been generated does not belong to you intellectually - it belongs to the model, and therefore to its owner, the AI company? Of course, there are explainers and assurances, but none of this is legally tested. As someone who has advised on, joined, and started lawsuits in more than a handful different legal systems, I can tell you that they are all very idiosyncratic and there is no common-sense to any of them. Local law can be extremely surprising and in fact one place's laws can contradict another place's laws. There are also secondary effects as well. For example, can you get public funding for such a project? Can you be recommended by government entities and various organizations? Can you use or contribute to such work as a university employee, researcher, or student? This last one is particularly important for a project that lives and dies by the ability of researchers to contribute to it. I may have simply missed it in the deluge, but I have found that there has been little to no discussion of the effect of LLM contributions on the copyright of GHC. Perhaps before entering this "Brave New World", we should see if the ice is too thin to walk on. Especially during war time, there has been a multitude of legal documents that were proof of citizenship to one area or another. Sometimes they looked like the blue British passport that is all the rage these days, sometimes they were a tattered piece of paper in an envelope of you were lucky, with nothing more to certify its authenticity than a stamp, and a block-letter name written by someone who barely ever held a pen, at an outpost in the deep forest. This may have been all that a person, coming from a country that may not even exist anymore due to changes in war, has to go on for legal recognition as a citizen of any country. Many times such documents were ad-hoc or provisional and there wasn't even a way to check them for authenticity or correctness, if the checking was being done by a person a thousand miles away from the document's origin and who is not even able to read the alphabet let alone the language. Very often such documents would not be recognized; it was as much dep endent on the politics of the day (does the Soviet Union recognize the Croatians?) as on whether you had a pack of cigarettes to share with the militia man patrolling the path you were on. A solid passport from a powerful country got you out of trouble. The legal system around LLM contributions is changing. Very little of it is proven and what has been proven so far shows that they enjoy no protection. For better or worse whatever imaginary lines on the ground we imagine and agree upon within our little tribe will be completely disregarded and obliterated by the machine of war that sets out the order as the legal system solidifies. To understand how this can lead to disappointment, just ask people in Finland whether they are nostalgic for their lost territory. I believe the integration of LLM based output into GHC is an issue that requires prudence far beyond "community consensus", "robust conversations", and "good proposals". It can clearly have ramifications of legal, compliance, civic, policy, and otherwise government-related nature, should be treated seriously as such, and therefore the project should source informed analysis and recommendation from those whose job it is to do so. This in particular means we should be hiring and listening to impartial professionals and not sounding off as developers who don't know the first thing about such matters. I say all this as someone who uses AI on the daily and has worked at a frontier AI lab that has been acquired by interests related to the US government. I think GHC is more important than being able to work on it with specific techniques. Everyone employing asbestos had the best intentions and could never have known where it would lead, but out of this period in history we have learned to be careful about employing new technology. Currently, the only people talking about this on this list are asbestos users and carpenters, and no one's thought to ask the doctors or physicists. With best regards, Damian _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Am Mo 03.08.2026 12:51 schrieb Moritz Angermann via ghc-devs:
A rather extreme take and thought experiment then is: Shouldn’t copyleft-licensed projects campaign for excessive LLM use in permissive-licensed projects, while being very strict about LLM use in their own codebase? This might later open up the claim that substantial portions of copyleft code ended up in other projects, and as such those would have to be relicensed under copyleft licenses, if these claims hold up in court. Therefore, if the idealized goal is to make every software free software, not using LLMs oneselves but promoting their use while hoping for favourable legal outcomes seems a viable although risky strategy.
Right. So now we know why there is so much pushing towards LLMs in various open-source projects: it’s about bringing them under the GPL. Conspiracy theory: here we come. Moritz is in fact a secret agent of Richard Stallman. 😉 People, please don’t dive into such lines of thought. This discussion has already been ten times as long as it should have been. All the best, Wolfgang
Repeating this in a top post, for visibility. I think the most productive work we can do is answering the following organizational questions: 1. How do we locate the right lawyers? 2. Can the Haskell Foundation even fund them? 3. Does this require a fundraiser of some sort? 4. Could the budgeting and/or fundraiser be joined by other communities and projects with the same issue? LLMs vs copyright is not a uniquely GHC thing. Maybe it's time to reach out to other projects under similar licenses? Other programming languages? Python, Rust, Postgres, PHP, Ruby, etc? Apache foundation? Most of the other large programming communities are corporate-owned, so this is our initial list of options. 5. At least some other projects have lawyers on staff, so maybe they are available to provide guidance to the larger community, if copyright is their area of specialization?
I have reached out (as I said I would) to the HF asking if they can provide information from a qualified expert when it comes to the facts of the law. I've also opened a discourse thread to gather legal question around this topic for this purpose: https://discourse.haskell.org/t/common-legal-questions-around-llms/14489 If people have concrete questions I would welcome them there. The HF has acknowledging that they received my message but hasn't said anything on the topic beyond that received my message (yet). But I'm pretty optimistic that we can make something happen there. It's plausible that matching our catalogue of questions to those asked and already answered by other projects will be good enough. If we think it's not good enough then my instinct is that any competent lawyer specialised on IP rights should be able to either answer our questions or be able to refer to one that can. Last but not least I think given the state of the law in regards to LLMs and IP rights I think it's best to think of this as a *risk assessment*. In practice many of those questions around LLMs are not quite yet settled law, and won't be for years to come. Andreas On 03/08/2026 14:37, damiancoding--- via ghc-devs wrote:
Repeating this in a top post, for visibility.
I think the most productive work we can do is answering the following organizational questions:
1. How do we locate the right lawyers? 2. Can the Haskell Foundation even fund them? 3. Does this require a fundraiser of some sort? 4. Could the budgeting and/or fundraiser be joined by other communities and projects with the same issue? LLMs vs copyright is not a uniquely GHC thing. Maybe it's time to reach out to other projects under similar licenses? Other programming languages? Python, Rust, Postgres, PHP, Ruby, etc? Apache foundation? Most of the other large programming communities are corporate-owned, so this is our initial list of options. 5. At least some other projects have lawyers on staff, so maybe they are available to provide guidance to the larger community, if copyright is their area of specialization? _______________________________________________ ghc-devs mailing list --ghc-devs@haskell.org To unsubscribe send an email toghc-devs-leave@haskell.org
I have reached out (as I said I would) to the HF
Great stuff, very helpful. Thank you!
our catalogue of questions
Maybe we could start building such a catalogue of questions.
any competent lawyer specialised on IP rights should be able to either answer our questions
Be very wary about this. In my experience it's easy for lawyers who appear competent - and even *are* competent - to absolutely bungle jobs. I think the main reason is people aren't good at estimating how competent a lawyer will be *in that particular situation*.
or be able to refer to one that can.
Watch out for references too. Lawyers usually don't keep in touch with other lawyers at a professional level other than in adversarial situations. Any references are usually going to be based on friendship. The only time lawyers associate with other lawyers based on specialization and quality of work is when they become partners (or otherwise work in the same law company). In my experience lawyer references from other lawyers have a high chance of failure.
risk assessment
agreed. So now the question would be: Why risk anything at all? Why risk decades of research and work by thousands of people? What are we winning that otherwise cannot be obtained at all? I think this needs assessment of its own, and I think it's super important. One common defense I've seen is "but then people will lie about their use of LLMs" - ok, and? If I'm a store clerk, and I'm given money, I'm not liable for where it came from, unless someone specifically tells me it came from a robbery. If people tell us they created things themselves, and they lied, then perhaps any copyright based quarrel would be possible to direct to that person (they lied and misled the project to accepting their derived, GPL-infected, stolen-IP code*) and deflect it off of GHC. * I'm not saying all LLM code is like that, those are just some possible arguments.
It's been a week - has there been any reply from the Haskell Foundation?
no On 8/11/2026 1:31 AM, damiancoding--- via ghc-devs wrote:
It's been a week - has there been any reply from the Haskell Foundation? _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
I've heard back from the HF in regards to facilitating legal advice/consultation and sadly the answer boiled down to no. On 11/08/2026 01:31, damiancoding--- via ghc-devs wrote:
It's been a week - has there been any reply from the Haskell Foundation? _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
What good are they then? Why does the HF exist? This might be the only real reason to have them around at all that we've had in 10 years, other than uh... hosting and domains?
Btw., the HF response can be seen here: https://groups.google.com/u/1/a/haskell.foundation/g/board/c/LoGTOAWiFmE On Wed, 2026-08-19 at 18:36 +0000, damiancoding--- via ghc-devs wrote:
What good are they then? Why does the HF exist? This might be the only real reason to have them around at all that we've had in 10 years, other than uh... hosting and domains? _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Let me copy the emails in here. First Andreas quoted my original email, then he wrote: Dear Haskell Foundation,
I'm contacting the HF as I think it would be helpful to get legal expertise through a legal consultation or similar to help inform AI policies and handling of LLM generated source contributions in various haskell projects.
In particular getting answers to these questions would be valuable:
==========================
Soundness of the "content right waivers" from popular LLM providers. These attempt to assign rights for generated content to users. Legal expertise about risks associated with above waivers, in particular due to clauses that might retroactively extinguish these rights due to terms of use violations would be welcome.
Consequences and potential risks when parts of a code base are not legally protected by copyright due non-copyrightable material. Are there unexpected consequences like changes to (lack of) liability for contributors, maintainers or organizations owning such projects? Especially for commonly used licenses in the haskell eco system (BSD3, MIT, (L)GPL, Apache2).
Risk assessment for becoming liable for copyright infringement due to accidental similarity in LLM output to copyrighted material.
Or any other legal pitfalls related to these subjects that we as legal laymen would be likely to miss.
==========================
I think it would be great if the HF could provide information and/or guidance on these issues driven by legal expertise. Please let me know if you think this could be arranged.
I've attached the discussion from which these questions arose below for additional context.
Best wishes Andreas
------------
Then, Laurent P. René de Cotret <lau...@haskell.foundation> wrote the email went straight to spam, and then he replied the following, copying in <Simon Peyton Jones>, <Ben Gamari>, and the user <board>
Thank you for your patience, Andreas.
Thank you for your patience, Andreas.
The Haskell Foundation Board discussed the issue, and the consensus is that dedicating HF resources to legal advice regarding LLMs is not a good use of our resources.
It's not that the questions are not important, but rather that this isn't Haskell-specific. If a larger body decides to get involved (e.g. the Free Software Foundation), the Haskell Foundation could collaborate as part of a larger coalition.
Best regards, Laurent ---------- Laurent P. René de Cotret Chair, Haskell Foundation
That's a pretty bad answer. Why does it matter if those questions are "Haskell specific"? We're talking about the specific impact on the GHC code base and its legal status. It doesn't matter if those are questions that also need to be answered by other projects, WE have to answer them for ourselves. That's like saying you don't need health insurance as an electrician because health insurance "isn't electrician specific". Sorry, but that's just plain nonsense and irresponsible. I question Laurent's legal instincts as a whole. The optics of this are pretty bad.
For what it's worth, I think it's a very good answer. The Haskell Foundation chooses how to spend limited resources. Their mandate is to spend those resources to fill in gaps in the Haskell ecosystem. Spending resources instead to obtain one more lawyer's opinion on issues that thousands of people are already thinking about a courts across the world are already actively grappling with is not nothing, but the incremental benefit is small. It's not clear how much difference one more lawyer would make in anyone's understanding of questions that are just not settled, at a level much broader than Haskell. I could see the answer being different if the original questions had brought up unique problems specific to GHC or major Haskell projects. I don't see any questions there, though, that differ from those facing pretty much every open source code base. On Thu, Aug 20, 2026 at 6:00 AM damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Let me copy the emails in here. First Andreas quoted my original email, then he wrote:
Dear Haskell Foundation,
I'm contacting the HF as I think it would be helpful to get legal
expertise through a legal consultation or similar
to help inform AI policies and handling of LLM generated source contributions in various haskell projects.
In particular getting answers to these questions would be valuable:
==========================
Soundness of the "content right waivers" from popular LLM providers. These attempt to assign rights for generated content to users. Legal expertise about risks associated with above waivers, in particular due to clauses that might retroactively extinguish these rights due to terms of use violations would be welcome.
Consequences and potential risks when parts of a code base are not legally protected by copyright due non-copyrightable material. Are there unexpected consequences like changes to (lack of) liability for contributors, maintainers or organizations owning such projects? Especially for commonly used licenses in the haskell eco system (BSD3, MIT, (L)GPL, Apache2).
Risk assessment for becoming liable for copyright infringement due to accidental similarity in LLM output to copyrighted material.
Or any other legal pitfalls related to these subjects that we as legal laymen would be likely to miss.
==========================
I think it would be great if the HF could provide information and/or guidance on these issues driven by legal expertise. Please let me know if you think this could be arranged.
I've attached the discussion from which these questions arose below for additional context.
Best wishes Andreas
------------
Then, Laurent P. René de Cotret <lau...@haskell.foundation> wrote the email went straight to spam, and then he replied the following, copying in <Simon Peyton Jones>, <Ben Gamari>, and the user <board>
Thank you for your patience, Andreas.
Thank you for your patience, Andreas.
The Haskell Foundation Board discussed the issue, and the consensus is that dedicating HF resources to legal advice regarding LLMs is not a good use of our resources.
It's not that the questions are not important, but rather that this isn't Haskell-specific. If a larger body decides to get involved (e.g. the Free Software Foundation), the Haskell Foundation could collaborate as part of a larger coalition.
Best regards, Laurent ---------- Laurent P. René de Cotret Chair, Haskell Foundation
That's a pretty bad answer. Why does it matter if those questions are "Haskell specific"? We're talking about the specific impact on the GHC code base and its legal status. It doesn't matter if those are questions that also need to be answered by other projects, WE have to answer them for ourselves. That's like saying you don't need health insurance as an electrician because health insurance "isn't electrician specific". Sorry, but that's just plain nonsense and irresponsible. I question Laurent's legal instincts as a whole. The optics of this are pretty bad. _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Losing control of the GHC code base would be a pretty big gap in the Haskell ecosystem. They don't have to go out and pay a lawyer out of pocket right away. There's myriad other ways, such as: - finding other projects affected by this, and seeing if they have already answered these questions - finding one of the "thousands of people" as you say which are "are already thinking" about those issues and asking them - banding together with them to pay for legal fees - asking the usual sponsors for extra money for this extraordinary need - raising funds in the community - running a gofundme Probably more, this is just off the top of my head.
the incremental benefit is small
So we're past laypeople theorizing what the legal impact of LLM use on the GHC code base could be, and straight to "any actual informed opinion by an expert will be of small benefit". That is very precarious thing to claim.
I could see the answer being different if the original questions had brought up unique problems specific to GHC or major Haskell projects
Once again: the fact that other people are on fire doesn't mean you shouldn't put out yourself.
Damian, You seem to be well versed in the legal field apparently. Can you help us understand how exactly we’d be losing control of the GHC codebase? Maybe if you could illustrate the attack vector, that would help us better understand the severity of the situation? As someone who has contributed to the GHC codebase quite a bit, even without LLMs, I am quite curious! Who, why and how would they take away the control of the GHC codebase? Best, Moritz On Thu, Aug 20, 2026 at 9:35 PM damiancoding--- via ghc-devs < ghc-devs@haskell.org> wrote:
Losing control of the GHC code base would be a pretty big gap in the Haskell ecosystem.
They don't have to go out and pay a lawyer out of pocket right away. There's myriad other ways, such as:
- finding other projects affected by this, and seeing if they have already answered these questions - finding one of the "thousands of people" as you say which are "are already thinking" about those issues and asking them - banding together with them to pay for legal fees - asking the usual sponsors for extra money for this extraordinary need - raising funds in the community - running a gofundme
Probably more, this is just off the top of my head.
the incremental benefit is small
So we're past laypeople theorizing what the legal impact of LLM use on the GHC code base could be, and straight to "any actual informed opinion by an expert will be of small benefit". That is very precarious thing to claim.
I could see the answer being different if the original questions had brought up unique problems specific to GHC or major Haskell projects
Once again: the fact that other people are on fire doesn't mean you shouldn't put out yourself. _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
There are a few ideas of how trouble could arise off the top of my head. None of this is going to be nearly as exhaustive as an expert legal opinion could be. You only need to look at the YouTube musician debacle for an example. Musicians on YouTube are completely unprotected and so people upload their music, then someone else copyrights it, and then claims the copyright. Shaky ownership of copyright of included code could easily result in a weak position when such challenges arise for GHC in the future. The tech industry has a history of companies making their whole existence out of suing thousands, hundreds of thousands of companies and people. "Piracy lawsuit" companies. Patent trolls. DMCA trolls. Etc. It is easy to imagine that in the future, a company could arise that: 1. Has identified that (some/every) AI training corpus included their copyrighted work 2. Deems any work generated by that AI to be derivative of their copyright 3. Harvests git repositories for people's emails from commits authored by John Doe "and claude" 4. Sends out one million fine / settlement demands Let's say they estimate the Haskell industry, which includes banks, high frequency traders, insurers, crypto, and other juicy steaks, to be worth $100M to them and that's their demand. What do we do then? Nothing, we're just in a deep hole, because Laurent de Cotret decided "this isn't Haskell-specific". You asked for an example. Here is one. I want to specifically preclude a conversation resulting from this that goes like: 1. "OK, I see your problem" 2. "But, aha! <some objection>" 3. "Therefore, there is no problem" 4. "So, we don't need experts looking into this." I will not be going down this road. Even if my simple example above which I came up with in 15 seconds of thinking is provably never going to happen, there can be other scenarios. This is just one thing to illustrate how things could go, but again, it's up to the experts to tell us what problems could *actually* arise.
On Thu, Aug 20, 2026 at 01:11:51PM -0000, damiancoding--- via ghc-devs wrote:
There are a few ideas of how trouble could arise off the top of my head. None of this is going to be nearly as exhaustive as an expert legal opinion could be. [...]
Is it fair to say that your concern arises from the possibility that code generate by an LLM might be legally judged (at some point in the future) to be a derived work of some part of its training data, and thus subject to a legal copyright claim from the copyright holder of that part of the training data? Tom
No, my concern arises from the fact that no lawyers have looked at this clearly complex legal situation. Please see the last part of my previous email. I do not want to get into taking apart an ad-hoc example where a holistic analysis of the whole legal landscape is necessary.
BTW - I wrote the original email a full 20 days before coming up with that simple illustrative example - and I should add I am not a Timelord
On Thu, Aug 20, 2026 at 02:06:33PM -0000, damiancoding--- via ghc-devs wrote:
No, my concern arises from the fact that no lawyers have looked at this clearly complex legal situation. Please see the last part of my previous email. I do not want to get into taking apart an ad-hoc example where a holistic analysis of the whole legal landscape is necessary.
If you're not concerned about copyright specifically, is it fair to say you are concerned about arbitrary legal issues in general that might arise, and you're not prepared to circumscribe an area of law to which those concerns belong? Tom
I am concerned about copyright, but not only, and I am specifically aware that I only have a very narrow perception compared to someone capable of doing systematic research in this space. I think it would be wise to look in the area of copyright, liability, as well as prospects in compliance with various forms of funding and income (such as government contracts). I wouldn't limit the search to these areas. Problems often come from unusual directions.
On Thu, Aug 20, 2026 at 09:26:16PM -0000, damiancoding--- via ghc-devs wrote:
I am concerned about copyright, but not only, and I am specifically aware that I only have a very narrow perception compared to someone capable of doing systematic research in this space. I think it would be wise to look in the area of copyright, liability, as well as prospects in compliance with various forms of funding and income (such as government contracts). I wouldn't limit the search to these areas. Problems often come from unusual directions.
And just so I'm 100% clear in my understanding of your point, can you confirm my guess that, as far as you know, every single open source project would be subject to the same sort of legal risks, if it were to incorporate LLM-generated code? (Of course, the impact of the risks may differ depending on the size and prominence of the project, who its users are and what they use it for.) Tom
Let's say they estimate the Haskell industry, which includes banks, high frequency traders, insurers, crypto, and other juicy steaks, to be worth $100M to them and that's their demand. What do we do then? Nothing, we're just in a deep hole, because Laurent de Cotret decided "this isn't Haskell-specific”.
Please don’t attack people. As you can see from the emails you quoted yourself, Laurent merely relayed the message of the Haskell Foundation Board. The decision was made by the board, not by Laurent alone; it shouldn’t be surprising that the boards chair is tasked with external communication.
I agree that we should maybe tone down a bit. But I also feel the HF response was fairly disappointing: - it says the topic is not Haskell specific, which is a bit confusing... is gitlab infrastructure not Haskell specific either? We're investing in all sorts of things that are not directly related to the language, but to the ecosystem at large. Protecting our "assets" legally is definitely one of them. - it says "not good use of our resources", but didn't actually provide insights as to why: e.g. is legal advice on that topic too expensive or is it a waste of money? Why? I mean it may very well be, but we don't have to get deep into crazy legal fees for a first consultation. - it ends with a hint that there are actually relevant action points (like... contacting the FSF), but then didn't seem to express any interest in following up on those. I would have expected the HF to take the lead here, because the success rate will be higher when it comes from such a body. We have to advertise our interest in this issue and not stand by passively. Cheers, Julian On Fri, 2026-08-21 at 12:03 +0900, Moritz Angermann via ghc-devs wrote:
Let's say they estimate the Haskell industry, which includes banks, high frequency traders, insurers, crypto, and other juicy steaks, to be worth $100M to them and that's their demand. What do we do then? Nothing, we're just in a deep hole, because Laurent de Cotret decided "this isn't Haskell-specific”.
Please don’t attack people. As you can see from the emails you quoted yourself, Laurent merely relayed the message of the Haskell Foundation Board. The decision was made by the board, not by Laurent alone; it shouldn’t be surprising that the boards chair is tasked with external communication.
_______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
It wasn't really a personal attack, it was a pure statement of fact and an attribution of (highly hypothetical) possible outcome. You can argue about the correct target for this attribution, but to me the director of the HF is the person responsible for anything the HF does and is the ultimate decision maker. I don't want to delve too much into the bylaws about that, the power dynamics ultimately depend less on the theoretical structure and more on everyone's personal actions. I am as baffled as Julian about the outcome, but I also admit perhaps no one in the HF even realizes this is a problem to be taken seriously. However if this position does not change I strongly suggest exploring possibilities for changing the board.
Hi Damian, To confirm what Moritz already said, the decision was made by the HF Board, not by the Chair alone. FWIW as one of directors of the HF, I support the decision. You are very welcome to nominate yourself when the next round of board elections will be happening. On the substance of matter, any meaningful legal advice from intellectual property lawyers on the subject is likely to cost five digits (unless the said lawyer will meat-proxy what ChatGPT says, of course). Given that hundreds of major companies all over the world use LLMs to write their cherished all-rights-reserved software (= their stakes to lose copyright are much more significant than for GHC team) and their lawyers are seemingly satisfied with it, there does not seem to be an immediate or visible danger. Ultimately at the moment all legal opinions on LLMs are a coin toss, until the matter is actually tested by courts. There is simply not enough legal evidence at the moment for lawyers to provide anything resembling a definitive answer. So to answer Julian's question above (in my private capacity and in the capacity of one of HF directors), my view is that a legal advice will be both expensive and a waste of time. Best regards, Andrew
On Sat, Aug 22, 2026 at 11:46:20AM -0000, andrew.lelechenko--- via ghc-devs wrote:
On the substance of matter, any meaningful legal advice from intellectual property lawyers on the subject is likely to cost five digits (unless the said lawyer will meat-proxy what ChatGPT says, of course). Given that hundreds of major companies all over the world use LLMs to write their cherished all-rights-reserved software (= their stakes to lose copyright are much more significant than for GHC team) and their lawyers are seemingly satisfied with it, there does not seem to be an immediate or visible danger. Ultimately at the moment all legal opinions on LLMs are a coin toss, until the matter is actually tested by courts. There is simply not enough legal evidence at the moment for lawyers to provide anything resembling a definitive answer.
FWIW, the OpenSSL project has an AI policy: https://openssl-library.org/post/2026-06-18-ai-policy/ https://openssl-library.org/policies/general/ai-policy/ The Two Core Requirements If a non-trivial portion of your submission has been created using an AI tool, you must: 1. declare it in your commit message using an Assisted-by trailer, and 2. have signed a Contributor License Agreement (CLA) that includes the AI clauses. Both requirements apply together. Neither alone is sufficient. -- Viktor. 🇺🇦 Слава Україні!
HI Viktor, Interesting. I thought about possible applications of CLAs when writing the previous message, but could not find a convincing case, where they would provide any additional guarantees for an open-source project. I wonder what's the reason behind OpenSSL doing it this way. Is it perhaps to prevent a malicious committer, who introduced a security vulnerability, from avoiding legal responsibility by saying "it was not me, it was all LLM's doing"?.. Best regards, Andrew
On Sat, Aug 22, 2026 at 05:18:41PM -0000, andrew.lelechenko--- via ghc-devs wrote:
Interesting. I thought about possible applications of CLAs when writing the previous message, but could not find a convincing case, where they would provide any additional guarantees for an open-source project. I wonder what's the reason behind OpenSSL doing it this way. Is it perhaps to prevent a malicious committer, who introduced a security vulnerability, from avoiding legal responsibility by saying "it was not me, it was all LLM's doing"?..
My perhaps naïve understanding of the purpose of the CLA is that it makes it possible for the project to adjust the copyright conditions as needed (just once so far in ~26 years) without having to seek the consent of every past contributor. Contributors backed by an employer need to file both an individual CLA and a corporate CLA. Copyright and liability are I would conjecture separate concerns, and in any case not the motivation for CLAs. -- Viktor. 🇺🇦 Слава Україні!
On 08/22/2026 1:46 PM CEST andrew.lelechenko--- via ghc-devs <ghc-devs@haskell.org> wrote: [...]
Without weighing in at all on the Haskell Foundation's decision, I must ask: don't these two adjacent sentences contradict one another?:
Given that hundreds of major companies all over the world use LLMs to write their cherished all-rights-reserved software (= their stakes to lose copyright are much more significant than for GHC team) and their lawyers are seemingly satisfied with it, there does not seem to be an immediate or visible danger.
Ultimately at the moment all legal opinions on LLMs are a coin toss, until the matter is actually tested by courts.
The fact that "hundreds of major companies" have made a risk-reward calculation and landed on the side of LLM use doesn't seem like very firm footing for a project that's invaluable and needs to be with us long after many of these companies have bitten the dust. Small companies may have decided they need to take the risk in order to keep up short-term profits; larger companies may have decided they have the army of lawyers and strategic relationships that make the risks worth taking. Neither of these strategies make good sense for GHC. Certainly hundreds of other companies have decided _against_ the use of LLMs on grounds of intellectual property and trade secret protections. To decide not to worry about GHC's copyright status because lawyers for other companies have made a decision -- a decision whose reasoning was hidden from us -- seems dangerously close to cargo-culting. Cheers, Tom
Hi Tom,
don't these two adjacent sentences contradict one another?
I don't see how. Is there "an immediate or visible danger"? No. Is there a way to eliminate any potential danger around LLM usage? No, "until the matter is actually tested by courts". Is it worth to spend five digits on some sort of "middle way", which might alleviate but not mitigate a potential but unspecified danger? Also no. Best regards, Andrew
this copyright thread confuses me, because …. what Haskell code thats public could these model echo into a ghc edit that isnt gonna be basically ghc code? the architecture and naming conventions and code style is quite distinct and theres very few salient hs code bases that would actually be salient, and translating an algorithm or data structure from one programming language does not create copyright risk wrt the translated code except to the extent string literals or specific data is copyrighted… i would love examples that ground the ip risk with now if we were talking abiut a bsd open source nosql db or web server, there’s would be lots to worry about agpl / gpl/ that commercial one that converts to apache after k years. in the case of ghc the very style and architecture just creates a risk of a model generating old no longer correct ancient ghc / haskell code! caveats: my family is basically all software ip lawyers and ive had to navigate quite complex ip nuances in my work over time. On Sun, Aug 23, 2026 at 2:34 AM andrew.lelechenko--- via ghc-devs < ghc-devs@haskell.org> wrote:
Hi Tom,
don't these two adjacent sentences contradict one another?
I don't see how. Is there "an immediate or visible danger"? No. Is there a way to eliminate any potential danger around LLM usage? No, "until the matter is actually tested by courts". Is it worth to spend five digits on some sort of "middle way", which might alleviate but not mitigate a potential but unspecified danger? Also no.
Best regards, Andrew _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
You seem to assume that cross-language copyright infringement is impossible and that copyright infringement must involve either very local things like naming conventions or global things like architecture. I am not learned in software IP law, but the above assumptions seem plainly false from the little I do know. It could just be a few functions from another project, possibly in a different programming language, adapted to Haskell and the GHC naming conventions. Also, there are plenty of compilers or just general libraries written in Haskell which contain code that could also be used in GHC. The fact that there is little comparable code in the training set could also mean the LLM is more likely to copy a significant part of a single project as there is less code to “learn” from. Cheers, Jaro On 9/1/26 17:26, Carter Schonwald via ghc-devs 'ghc-devs at haskell.org' wrote:
this copyright thread confuses me, because …. what Haskell code thats public could these model echo into a ghc edit that isnt gonna be basically ghc code? the architecture and naming conventions and code style is quite distinct and theres very few salient hs code bases that would actually be salient, and translating an algorithm or data structure from one programming language does not create copyright risk wrt the translated code except to the extent string literals or specific data is copyrighted…
i would love examples that ground the ip risk with
now if we were talking abiut a bsd open source nosql db or web server, there’s would be lots to worry about agpl / gpl/ that commercial one that converts to apache after k years. in the case of ghc the very style and architecture just creates a risk of a model generating old no longer correct ancient ghc / haskell code!
caveats: my family is basically all software ip lawyers and ive had to navigate quite complex ip nuances in my work over time.
On Sun, Aug 23, 2026 at 2:34 AM andrew.lelechenko--- via ghc-devs <ghc- devs@haskell.org <mailto:ghc-devs@haskell.org>> wrote:
Hi Tom,
> don't these two adjacent sentences contradict one another?
I don't see how. Is there "an immediate or visible danger"? No. Is there a way to eliminate any potential danger around LLM usage? No, "until the matter is actually tested by courts". Is it worth to spend five digits on some sort of "middle way", which might alleviate but not mitigate a potential but unspecified danger? Also no.
Best regards, Andrew _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org <mailto:ghc-devs@haskell.org> To unsubscribe send an email to ghc-devs-leave@haskell.org <mailto:ghc- devs-leave@haskell.org>
By the way, if the FSF is an authority on this matter, one could look at the policy of GCC, which is under the FSF's patronage: https://gcc.gnu.org/ai-policy.html Obv GCC has different license type and they're much more serious about copyright than GHC ever did, so opinion from a qualified person would be valuable there. On 8/20/26 16:00, damiancoding--- via ghc-devs wrote:
Let me copy the emails in here. First Andreas quoted my original email, then he wrote:
Dear Haskell Foundation,
I'm contacting the HF as I think it would be helpful to get legal expertise through a legal consultation or similar to help inform AI policies and handling of LLM generated source contributions in various haskell projects.
In particular getting answers to these questions would be valuable:
==========================
Soundness of the "content right waivers" from popular LLM providers. These attempt to assign rights for generated content to users. Legal expertise about risks associated with above waivers, in particular due to clauses that might retroactively extinguish these rights due to terms of use violations would be welcome.
Consequences and potential risks when parts of a code base are not legally protected by copyright due non-copyrightable material. Are there unexpected consequences like changes to (lack of) liability for contributors, maintainers or organizations owning such projects? Especially for commonly used licenses in the haskell eco system (BSD3, MIT, (L)GPL, Apache2).
Risk assessment for becoming liable for copyright infringement due to accidental similarity in LLM output to copyrighted material.
Or any other legal pitfalls related to these subjects that we as legal laymen would be likely to miss.
==========================
I think it would be great if the HF could provide information and/or guidance on these issues driven by legal expertise. Please let me know if you think this could be arranged.
I've attached the discussion from which these questions arose below for additional context.
Best wishes Andreas
------------ Then, Laurent P. René de Cotret <lau...@haskell.foundation> wrote the email went straight to spam, and then he replied the following, copying in <Simon Peyton Jones>, <Ben Gamari>, and the user <board>
Thank you for your patience, Andreas.
Thank you for your patience, Andreas.
The Haskell Foundation Board discussed the issue, and the consensus is that dedicating HF resources to legal advice regarding LLMs is not a good use of our resources.
It's not that the questions are not important, but rather that this isn't Haskell-specific. If a larger body decides to get involved (e.g. the Free Software Foundation), the Haskell Foundation could collaborate as part of a larger coalition.
Best regards, Laurent ---------- Laurent P. René de Cotret Chair, Haskell Foundation That's a pretty bad answer. Why does it matter if those questions are "Haskell specific"? We're talking about the specific impact on the GHC code base and its legal status. It doesn't matter if those are questions that also need to be answered by other projects, WE have to answer them for ourselves. That's like saying you don't need health insurance as an electrician because health insurance "isn't electrician specific". Sorry, but that's just plain nonsense and irresponsible. I question Laurent's legal instincts as a whole. The optics of this are pretty bad.
ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Given that AI companies are now watermarking their outputs, the copyright discussion has become even more important. I would ask everyone to think about the implications of this - it feels like we're singing and dancing while the ship is about to hit an iceberg. We cannot be oblivious to what is happening right in front of us.
Am Mi 19.08.2026 18:39 schrieb damiancoding--- via ghc-devs:
Given that AI companies are now watermarking their outputs, the copyright discussion has become even more important. I would ask everyone to think about the implications of this - it feels like we’re singing and dancing while the ship is about to hit an iceberg. We cannot be oblivious to what is happening right in front of us.
The solution is simple: Just don’t add LLM-generated content to the GHC codebase. This would also be good for other reasons, as has been discussed up and down on this mailing list. All the best, Wolfgang
There seems to be controversy around such a solution. Given that the HF is unwilling to explore any method of gaining legal clarity, perhaps LLM proponents would like to sponsor a pilot project?
participants (14)
-
amindfv＠mailbox.org -
Andreas Klebinger -
Andrei Borzenkov -
andrew.lelechenko＠gmail.com -
Carter Schonwald -
Chris Smith -
code＠jaro.addy.io -
damiancoding＠gmail.com -
Hécate -
Julian Ospald -
Moritz Angermann -
Tom Ellis -
Viktor Dukhovni -
Wolfgang Jeltsch