A shared statement on our environment in the face of LLMs
Dear friends, I've spent the last two days discussing the topic of LLM contributions in detail with Julian at length over video. As the discussion over the last few weeks have shown that it's very hard to find agreement on a common policy, we have tried to outline what we consider a statement that covers at least both of our views. With Julian and I representing a fairly wide spectrum of the positions, we hope that this may provide a statement that many can subscribe to. We have tried to keep it short, and not get distracted by specificies of workflows, quality standards, and definitions of terms, which seem to distract from the core that we want to address. Respectfully, Julian and Moritz ## A shared statement on our environment in the face of LLMs As human-LLM interaction is becoming more common, we find ourselves debating not only our technical goals, but also our understanding of open source community and values. This is why we all agree that we want to participate in an environment... 1. that inspires the exchange of technical ideas and knowledge between humans 2. that is supportive and accessible to the point that no one should feel the need to use LLM in the first place 3. that facilitates building trust relationships with one another 4. that rewards technical excellence and authenticity 5. that demands all interactions to be genuine and respectful We recognize that using LLMs responsibly is challenging. Not only can they be used for abuse[^1][^2], but they may also cause the quality of communication and intellectual engagement to decline[^3]. There is an emerging body of scientific literature supporting these concerns about negative effects on psychological, cognitive, real-world performance, etc. We recommend everyone to familiarize themselves with the current state of research. Here are a few links for your consideration. * [How AI Impacts Skill Formation](https://arxiv.org/abs/2601.20245) * [AI Assistance Reduces Persistence and Hurts Independent Performance](https://arxiv.org/abs/2604.04721) * [From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms for Dignified Human-AI Interaction](https://arxiv.org/abs/2601.21920) * [AI Competency Erosion: Understanding Expertise Decay](https://link.springer.com/chapter/10.1007/978-3-032-11748-9_5) * [Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task](https://arxiv.org/abs/2506.08872) * [Investigating the Effects of LLM Use on Critical Thinking Under Time Constraints: Access Timing and Time Availability](https://arxiv.org/abs/2603.08849) _(This list does not claim to be comprehensive; it is merely meant to provide you with an initial stepping stone into the current research)._ Moreover we are especially concerned about their potential negative effects on open source communities[^4], human communication and collaboration. We strive to be an inclusive community and judge contributions on their merits. This includes the use of LLMs as long as it does not disturb our envisioned environment outlined earlier. We recognize that there are users and contributors of this project who would rather not interact with LLM generated content at all. To balance the preferences and self-determination of the diverse group of contributors, we require that the use of LLMs to generate content is clearly attributed as such, failure to do so may result in contributions to be rejected. We understand that this may not be satisfactory to everyone, since we cannot completely avoid exposure to LLM generated content. However, we will neither compromise our quality standards, nor allow any disruption of human-human interaction by the use of technology. Examples of inappropriate use are: - feeding someone's email into an LLM to generate an answer - creating automated LLMs written bug reports that a volunteer has to work through to determine if it has any value at all - generating large patches that take more time to review by the project than it took you to create We want this to be a project of human interaction, of shared curiosity and a beacon of engineering excellence. These are the shared values we stand by today as a project, and they may evolve along with our shared experiences and emerging scientific evidence over time. Julian Ospald Moritz Angermann -- [^1]: https://openai.com/index/disrupting-malicious-ai-uses/ [^2]: https://www.infoq.com/news/2026/02/ai-floods-close-projects/ [^3]: https://arxiv.org/abs/2506.06166 [^4]: https://arxiv.org/abs/2601.15494
Hi Moritz & Julian, I find this statement to be a good policy to adopt for the GHC project, since it puts the emphasis on human collaboration and fulfilment. Thanks for your work, both of you. Cheers, Hécate Le 28/07/2026 à 12:14, Moritz Angermann via ghc-devs a écrit :
Dear friends,
I've spent the last two days discussing the topic of LLM contributions in detail with Julian at length over video. As the discussion over the last few weeks have shown that it's very hard to find agreement on a common policy, we have tried to outline what we consider a statement that covers at least both of our views. With Julian and I representing a fairly wide spectrum of the positions, we hope that this may provide a statement that many can subscribe to.
We have tried to keep it short, and not get distracted by specificies of workflows, quality standards, and definitions of terms, which seem to distract from the core that we want to address.
Respectfully, Julian and Moritz
## A shared statement on our environment in the face of LLMs
As human-LLM interaction is becoming more common, we find ourselves debating not only our technical goals, but also our understanding of open source community and values.
This is why we all agree that we want to participate in an environment...
1. that inspires the exchange of technical ideas and knowledge between humans 2. that is supportive and accessible to the point that no one should feel the need to use LLM in the first place 3. that facilitates building trust relationships with one another 4. that rewards technical excellence and authenticity 5. that demands all interactions to be genuine and respectful
We recognize that using LLMs responsibly is challenging. Not only can they be used for abuse[^1][^2], but they may also cause the quality of communication and intellectual engagement to decline[^3].
There is an emerging body of scientific literature supporting these concerns about negative effects on psychological, cognitive, real-world performance, etc. We recommend everyone to familiarize themselves with the current state of research. Here are a few links for your consideration.
* [How AI Impacts Skill Formation](https://arxiv.org/abs/2601.20245) * [AI Assistance Reduces Persistence and Hurts Independent Performance](https://arxiv.org/abs/2604.04721) * [From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms for Dignified Human-AI Interaction](https://arxiv.org/abs/2601.21920) * [AI Competency Erosion: Understanding Expertise Decay](https://link.springer.com/chapter/10.1007/978-3-032-11748-9_5) * [Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task](https://arxiv.org/abs/2506.08872) * [Investigating the Effects of LLM Use on Critical Thinking Under Time Constraints: Access Timing and Time Availability](https://arxiv.org/abs/2603.08849)
_(This list does not claim to be comprehensive; it is merely meant to provide you with an initial stepping stone into the current research)._
Moreover we are especially concerned about their potential negative effects on open source communities[^4], human communication and collaboration.
We strive to be an inclusive community and judge contributions on their merits. This includes the use of LLMs as long as it does not disturb our envisioned environment outlined earlier.
We recognize that there are users and contributors of this project who would rather not interact with LLM generated content at all. To balance the preferences and self-determination of the diverse group of contributors, we require that the use of LLMs to generate content is clearly attributed as such, failure to do so may result in contributions to be rejected. We understand that this may not be satisfactory to everyone, since we cannot completely avoid exposure to LLM generated content. However, we will neither compromise our quality standards, nor allow any disruption of human-human interaction by the use of technology.
Examples of inappropriate use are:
- feeding someone's email into an LLM to generate an answer - creating automated LLMs written bug reports that a volunteer has to work through to determine if it has any value at all - generating large patches that take more time to review by the project than it took you to create
We want this to be a project of human interaction, of shared curiosity and a beacon of engineering excellence.
These are the shared values we stand by today as a project, and they may evolve along with our shared experiences and emerging scientific evidence over time.
Julian Ospald Moritz Angermann
-- [^1]:https://openai.com/index/disrupting-malicious-ai-uses/ [^2]:https://www.infoq.com/news/2026/02/ai-floods-close-projects/ [^3]:https://arxiv.org/abs/2506.06166 [^4]:https://arxiv.org/abs/2601.15494
_______________________________________________ ghc-devs mailing list --ghc-devs@haskell.org To unsubscribe send an email toghc-devs-leave@haskell.org
-- Hécate ✨ WWW:https://glitchbra.in RUN: BSD
Dear Moritz and Julian, thanks a lot to both of you for getting together, discussing, and ultimately crafting this statement. It matches my own views to a large extent. All the best, Wolfgang Am Di 28.07.2026 17:14 schrieb Moritz Angermann via ghc-devs:
Dear friends,
I've spent the last two days discussing the topic of LLM contributions in detail with Julian at length over video. As the discussion over the last few weeks have shown that it's very hard to find agreement on a common policy, we have tried to outline what we consider a statement that covers at least both of our views. With Julian and I representing a fairly wide spectrum of the positions, we hope that this may provide a statement that many can subscribe to.
We have tried to keep it short, and not get distracted by specificies of workflows, quality standards, and definitions of terms, which seem to distract from the core that we want to address.
Respectfully, Julian and Moritz
## A shared statement on our environment in the face of LLMs
As human-LLM interaction is becoming more common, we find ourselves debating not only our technical goals, but also our understanding of open source community and values.
This is why we all agree that we want to participate in an environment...
1. that inspires the exchange of technical ideas and knowledge between humans 2. that is supportive and accessible to the point that no one should feel the need to use LLM in the first place 3. that facilitates building trust relationships with one another 4. that rewards technical excellence and authenticity 5. that demands all interactions to be genuine and respectful
We recognize that using LLMs responsibly is challenging. Not only can they be used for abuse[^1][^2], but they may also cause the quality of communication and intellectual engagement to decline[^3].
There is an emerging body of scientific literature supporting these concerns about negative effects on psychological, cognitive, real-world performance, etc. We recommend everyone to familiarize themselves with the current state of research. Here are a few links for your consideration.
* [How AI Impacts Skill Formation](https://arxiv.org/abs/2601.20245) * [AI Assistance Reduces Persistence and Hurts Independent Performance](https://arxiv.org/abs/2604.04721) * [From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms for Dignified Human-AI Interaction](https://arxiv.org/abs/2601.21920) * [AI Competency Erosion: Understanding Expertise Decay](https://link.springer.com/chapter/10.1007/978-3-032-11748-9_5) * [Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task](https://arxiv.org/abs/2506.08872) * [Investigating the Effects of LLM Use on Critical Thinking Under Time Constraints: Access Timing and Time Availability](https://arxiv.org/abs/2603.08849)
_(This list does not claim to be comprehensive; it is merely meant to provide you with an initial stepping stone into the current research)._
Moreover we are especially concerned about their potential negative effects on open source communities[^4], human communication and collaboration.
We strive to be an inclusive community and judge contributions on their merits. This includes the use of LLMs as long as it does not disturb our envisioned environment outlined earlier.
We recognize that there are users and contributors of this project who would rather not interact with LLM generated content at all. To balance the preferences and self-determination of the diverse group of contributors, we require that the use of LLMs to generate content is clearly attributed as such, failure to do so may result in contributions to be rejected. We understand that this may not be satisfactory to everyone, since we cannot completely avoid exposure to LLM generated content. However, we will neither compromise our quality standards, nor allow any disruption of human-human interaction by the use of technology.
Examples of inappropriate use are:
- feeding someone's email into an LLM to generate an answer - creating automated LLMs written bug reports that a volunteer has to work through to determine if it has any value at all - generating large patches that take more time to review by the project than it took you to create
We want this to be a project of human interaction, of shared curiosity and a beacon of engineering excellence.
These are the shared values we stand by today as a project, and they may evolve along with our shared experiences and emerging scientific evidence over time.
Julian Ospald Moritz Angermann
-- [^1]: https://openai.com/index/disrupting-malicious-ai-uses/ [^2]: https://www.infoq.com/news/2026/02/ai-floods-close-projects/ [^3]: https://arxiv.org/abs/2506.06166 [^4]: https://arxiv.org/abs/2601.15494
Hi Moritz and Julian, Are you proposing this as an alternative to Simon's draft policy? Cheers, Teo On Tue, 28 Jul 2026, at 11:14 AM, Moritz Angermann via ghc-devs wrote:
Dear friends,
I've spent the last two days discussing the topic of LLM contributions in detail with Julian at length over video. As the discussion over the last few weeks have shown that it's very hard to find agreement on a common policy, we have tried to outline what we consider a statement that covers at least both of our views. With Julian and I representing a fairly wide spectrum of the positions, we hope that this may provide a statement that many can subscribe to.
We have tried to keep it short, and not get distracted by specificies of workflows, quality standards, and definitions of terms, which seem to distract from the core that we want to address.
Respectfully, Julian and Moritz
## A shared statement on our environment in the face of LLMs
As human-LLM interaction is becoming more common, we find ourselves debating not only our technical goals, but also our understanding of open source community and values.
This is why we all agree that we want to participate in an environment...
1. that inspires the exchange of technical ideas and knowledge between humans 2. that is supportive and accessible to the point that no one should feel the need to use LLM in the first place 3. that facilitates building trust relationships with one another 4. that rewards technical excellence and authenticity 5. that demands all interactions to be genuine and respectful
We recognize that using LLMs responsibly is challenging. Not only can they be used for abuse[^1][^2], but they may also cause the quality of communication and intellectual engagement to decline[^3].
There is an emerging body of scientific literature supporting these concerns about negative effects on psychological, cognitive, real-world performance, etc. We recommend everyone to familiarize themselves with the current state of research. Here are a few links for your consideration.
* [How AI Impacts Skill Formation](https://arxiv.org/abs/2601.20245) * [AI Assistance Reduces Persistence and Hurts Independent Performance](https://arxiv.org/abs/2604.04721) * [From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms for Dignified Human-AI Interaction](https://arxiv.org/abs/2601.21920) * [AI Competency Erosion: Understanding Expertise Decay](https://link.springer.com/chapter/10.1007/978-3-032-11748-9_5) * [Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task](https://arxiv.org/abs/2506.08872) * [Investigating the Effects of LLM Use on Critical Thinking Under Time Constraints: Access Timing and Time Availability](https://arxiv.org/abs/2603.08849)
_(This list does not claim to be comprehensive; it is merely meant to provide you with an initial stepping stone into the current research)._
Moreover we are especially concerned about their potential negative effects on open source communities[^4], human communication and collaboration.
We strive to be an inclusive community and judge contributions on their merits. This includes the use of LLMs as long as it does not disturb our envisioned environment outlined earlier.
We recognize that there are users and contributors of this project who would rather not interact with LLM generated content at all. To balance the preferences and self-determination of the diverse group of contributors, we require that the use of LLMs to generate content is clearly attributed as such, failure to do so may result in contributions to be rejected. We understand that this may not be satisfactory to everyone, since we cannot completely avoid exposure to LLM generated content. However, we will neither compromise our quality standards, nor allow any disruption of human-human interaction by the use of technology.
Examples of inappropriate use are:
- feeding someone's email into an LLM to generate an answer - creating automated LLMs written bug reports that a volunteer has to work through to determine if it has any value at all - generating large patches that take more time to review by the project than it took you to create
We want this to be a project of human interaction, of shared curiosity and a beacon of engineering excellence.
These are the shared values we stand by today as a project, and they may evolve along with our shared experiences and emerging scientific evidence over time.
Julian Ospald Moritz Angermann
-- [^1]: https://openai.com/index/disrupting-malicious-ai-uses/ [^2]: https://www.infoq.com/news/2026/02/ai-floods-close-projects/ [^3]: https://arxiv.org/abs/2506.06166 [^4]: https://arxiv.org/abs/2601.15494 _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Hi Teo, Simon's draft policy is what inspired and motivated the statement. I see this exactly as more of a statement and less of a policy. The only "policy" thing in our text is indeed "you need to label/attribute LLM generated content". And we mention that only because it is part of "protecting the no-LLM" side. I believe that the rest of Simon's policy is maybe a worthwhile effort, but is quite separate from the matter of "here's our position on LLM use". Quality standards, workflow expectations, notions of authorship etc. are indeed things that need to be defined separately, whether we ban or allow LLMs. So maybe this is the way forward: users who want to know where the project stands in the "LLM discussion" can read this text. For more specific details on quality requirements, etc. we need a "contribution policy" that lays out requirements on ownership, review expectations and so on. They can certainly link to one another. But my feeling is that we got sidetracked from the essence of our disagreement while descending into the depths of policy specifications. Cheers, Julian On Tue, 2026-07-28 at 21:04 +0100, Teo Camarasu wrote:
Hi Moritz and Julian,
Are you proposing this as an alternative to Simon's draft policy?
Cheers, Teo
On Tue, 28 Jul 2026, at 11:14 AM, Moritz Angermann via ghc-devs wrote:
Dear friends,
I've spent the last two days discussing the topic of LLM contributions in detail with Julian at length over video. As the discussion over the last few weeks have shown that it's very hard to find agreement on a common policy, we have tried to outline what we consider a statement that covers at least both of our views. With Julian and I representing a fairly wide spectrum of the positions, we hope that this may provide a statement that many can subscribe to.
We have tried to keep it short, and not get distracted by specificies of workflows, quality standards, and definitions of terms, which seem to distract from the core that we want to address.
Respectfully, Julian and Moritz
## A shared statement on our environment in the face of LLMs
As human-LLM interaction is becoming more common, we find ourselves debating not only our technical goals, but also our understanding of open source community and values.
This is why we all agree that we want to participate in an environment...
1. that inspires the exchange of technical ideas and knowledge between humans 2. that is supportive and accessible to the point that no one should feel the need to use LLM in the first place 3. that facilitates building trust relationships with one another 4. that rewards technical excellence and authenticity 5. that demands all interactions to be genuine and respectful
We recognize that using LLMs responsibly is challenging. Not only can they be used for abuse[^1][^2], but they may also cause the quality of communication and intellectual engagement to decline[^3].
There is an emerging body of scientific literature supporting these concerns about negative effects on psychological, cognitive, real- world performance, etc. We recommend everyone to familiarize themselves with the current state of research. Here are a few links for your consideration.
* [How AI Impacts Skill Formation](https://arxiv.org/abs/2601.20245) * [AI Assistance Reduces Persistence and Hurts Independent Performance](https://arxiv.org/abs/2604.04721) * [From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms for Dignified Human-AI Interaction](https://arxiv.org/abs/2601.21920) * [AI Competency Erosion: Understanding Expertise Decay](https://link.springer.com/chapter/10.1007/978-3-032-11748-9_ 5) * [Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task](https://arxiv.org/abs/2506.08872) * [Investigating the Effects of LLM Use on Critical Thinking Under Time Constraints: Access Timing and Time Availability](https://arxiv.org/abs/2603.08849)
_(This list does not claim to be comprehensive; it is merely meant to provide you with an initial stepping stone into the current research)._
Moreover we are especially concerned about their potential negative effects on open source communities[^4], human communication and collaboration.
We strive to be an inclusive community and judge contributions on their merits. This includes the use of LLMs as long as it does not disturb our envisioned environment outlined earlier.
We recognize that there are users and contributors of this project who would rather not interact with LLM generated content at all. To balance the preferences and self-determination of the diverse group of contributors, we require that the use of LLMs to generate content is clearly attributed as such, failure to do so may result in contributions to be rejected. We understand that this may not be satisfactory to everyone, since we cannot completely avoid exposure to LLM generated content. However, we will neither compromise our quality standards, nor allow any disruption of human-human interaction by the use of technology.
Examples of inappropriate use are:
- feeding someone's email into an LLM to generate an answer - creating automated LLMs written bug reports that a volunteer has to work through to determine if it has any value at all - generating large patches that take more time to review by the project than it took you to create
We want this to be a project of human interaction, of shared curiosity and a beacon of engineering excellence.
These are the shared values we stand by today as a project, and they may evolve along with our shared experiences and emerging scientific evidence over time.
Julian Ospald Moritz Angermann
-- [^1]: https://openai.com/index/disrupting-malicious-ai-uses/ [^2]: https://www.infoq.com/news/2026/02/ai-floods-close-projects/ [^3]: https://arxiv.org/abs/2506.06166 [^4]: https://arxiv.org/abs/2601.15494 _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
Am Mi 29.07.2026 11:44 schrieb Julian Ospald via ghc-devs:
But my feeling is that we got sidetracked from the essence of our disagreement while descending into the depths of policy specifications.
Exactly. And therefore I have the following idea: We could start with the shared statement and perhaps install a concrete policy later. This approach would acknowledge that we haven’t reached agreement on various concrete points, and it could shape any discussion about these. With the shared statement in place, we could also make any policy crisper, as we could focus on a few, simple rules instead of giving rules and also the motivation for these rules. What are your thoughts on this? All the best, Wolfgang
Hi folks, I would be strongly opposed to adopting this shared statement instead of a policy. The way I read it, this shared statement is actually an extremely weak document. It mentions some of the harms from LLMs but many of the negative aspects from Simon's draft are completely dropped. It also fails to connect any of these harms to its very weak policy suggestions. This is also a significant step back from the existing policy <https://gitlab.haskell.org/ghc/ghc/-/wikis/contributing/AI>. This policy specifies that:
Contributors should only put forth contributions that they could have created without the support of AI tooling. To drop this completely would be a large step back.
I'm also very confused about the notion that attribution is a way of "protecting" the no-LLM side. While some people might hold this view, it doesn't seem like the dominant position I've seen in this conversation so far. The main thing I've seen is that people want attribution because it allows them to review LLM generated code with an extra keen eye. I think it also doesn't do a very good job of portraying the anti-LLM position in that paragraph. While some people merely dislike seeing LLM output, for others there are stronger objections to LLM use and the tone comes could do with improvement. Is the intention that if a user does one of the things in "inappropriate use" that they're contribution would be rejected? These seem like extremely controversial suggestions to me, and also things that are very difficult to tell. My impression is that Simon's draft while not satisfying everyone was close to being ready to be adopted. A great deal of time and debate has gone into making it. Adopting a weak statement and kicking the difficult task of figuring out the policy (which is mostly done!) down the road is unlikely to make this task easier in the future. I think this long drawn out process has already alienated a lot of people from engaging in discussion, and restarting the process from scratch doesn't sound like a wise idea to me. Cheers, Teo On Wed, 29 Jul 2026, at 10:34 AM, Wolfgang Jeltsch wrote:
Am Mi 29.07.2026 11:44 schrieb Julian Ospald via ghc-devs:
But my feeling is that we got sidetracked from the essence of our disagreement while descending into the depths of policy specifications.
Exactly. And therefore I have the following idea: We could start with the shared statement and perhaps install a concrete policy later. This approach would acknowledge that we haven’t reached agreement on various concrete points, and it could shape any discussion about these. With the shared statement in place, we could also make any policy crisper, as we could focus on a few, simple rules instead of giving rules and also the motivation for these rules.
What are your thoughts on this?
All the best, Wolfgang
On Wed, 2026-07-29 at 10:54 +0100, Teo Camarasu via ghc-devs wrote:
Hi folks,
I would be strongly opposed to adopting this shared statement instead of a policy.
This was not the suggestion. I think I explained very clearly that it is not a substitute for a "contribution policy". But for Moritz and myself it does a better job at capturing the opposing positions, which I do not think the current policy draft delivers. I don't have a strong opinion on what the GHC project does or does not do. It's more likely we will adopt this statement for ourselves. But we felt it was worth sharing. Cheers, Julian
Hi Julian, Sorry for the confusion! I was replying to Wolfgang's suggestion that we adopt this statement first and then come up with a policy later. I think this document does a good job of being a statement and especially as a statement resolving your disagreements! I agree with a lot of what's in it. Thanks for sharing it. Cheers, Teo
Hi, Teo! I completely understand your concerns and their justifications. That said, an advantage of the shared statement over the draft policy is, in my opinion, that the shared statement stresses jointly-held, positive values more and more strongly points out the risks of LLM use with regards to these values. Yes, you could say that it “allows” more than the draft policy, but this is because it’s not a policy and therefore doesn’t regulate the details. When respecting the message of the shared statement, one should find that lots of things around LLMs are rendered non-permissible or at least highly questionable by it. Also, don’t forget that in the end the draft policy didn’t have so much support from the “anti-LLM” side anymore, because it had been watered down. Maybe, I’m not seeing this issue clearly. I must say that I can’t afford anymore spending time to think more deeply about all those policy proposals and statements and I’m actually pretty sick of this whole discussion. An advantage of the shared statement for me is that it seems clear to me what its message is. The draft policy, with its detours into various facets, rather confuses me. All the best, Wolfgang Am Mi 29.07.2026 10:54 schrieb Teo Camarasu:
Hi folks,
I would be strongly opposed to adopting this shared statement instead of a policy.
The way I read it, this shared statement is actually an extremely weak document. It mentions some of the harms from LLMs but many of the negative aspects from Simon’s draft are completely dropped. It also fails to connect any of these harms to its very weak policy suggestions.
This is also a significant step back from the existing policy <https://gitlab.haskell.org/ghc/ghc/-/wikis/contributing/AI>. This policy specifies that:
Contributors should only put forth contributions that they could have created without the support of AI tooling.
To drop this completely would be a large step back.
I’m also very confused about the notion that attribution is a way of “protecting” the no-LLM side. While some people might hold this view, it doesn’t seem like the dominant position I’ve seen in this conversation so far. The main thing I’ve seen is that people want attribution because it allows them to review LLM generated code with an extra keen eye.
I think it also doesn’t do a very good job of portraying the anti-LLM position in that paragraph. While some people merely dislike seeing LLM output, for others there are stronger objections to LLM use and the tone comes could do with improvement.
Is the intention that if a user does one of the things in “inappropriate use” that they’re contribution would be rejected? These seem like extremely controversial suggestions to me, and also things that are very difficult to tell.
My impression is that Simon’s draft while not satisfying everyone was close to being ready to be adopted. A great deal of time and debate has gone into making it.
Adopting a weak statement and kicking the difficult task of figuring out the policy (which is mostly done!) down the road is unlikely to make this task easier in the future. I think this long drawn out process has already alienated a lot of people from engaging in discussion, and restarting the process from scratch doesn’t sound like a wise idea to me.
Cheers, Teo
On Wed, 29 Jul 2026, at 10:34 AM, Wolfgang Jeltsch wrote:
Am Mi 29.07.2026 11:44 schrieb Julian Ospald via ghc-devs:
But my feeling is that we got sidetracked from the essence of our disagreement while descending into the depths of policy specifications.
Exactly. And therefore I have the following idea: We could start with the shared statement and perhaps install a concrete policy later. This approach would acknowledge that we haven’t reached agreement on various concrete points, and it could shape any discussion about these. With the shared statement in place, we could also make any policy crisper, as we could focus on a few, simple rules instead of giving rules and also the motivation for these rules.
What are your thoughts on this?
All the best, Wolfgang
Hi Teo,
Contributors should only put forth contributions that they could have created without the support of AI tooling.
I agree with that statement. Although I'd probably add reasonably to it. I do think it suffers from the same "very difficult to tell":
Is the intention that if a user does one of the things in "inappropriate use" that they're contribution would be rejected? These seem like extremely controversial suggestions to me, and also things that are very difficult to tell.
Out of curiosity, which of the inappropriate use cases listed do you find extremely controversial? I'm still operating under the premise that there is neither a right to have one's contribution reviewed, nor a right to have one's contribution included/merged. It's ultimately up to the core maintainers/folks with commit bits to decide if they deem a contribution inclusion worthy or not. Maybe this assumption is fundamentally flawed and just wrong? Best, Moritz
Hey folks, I apologize that my only other contribution to this discussion was a little emotional. Harendra wrote something that really resonated with me (positively)! Moritz and Julian's shared statement is one I fully support. I also support the draft policy. I agree that such a policy is needed. I note (with satisfaction) that you could just as easily call it a general "Contribution Policy", but that's neither here nor there. The main takeaway is that GHC is a project designed for human eyeballs. I think we have 100% agreement on that particular value? The one thing I reacted to previously was P3: declaring use of LLMs. Now, I realize it's literally the least I could do. Of course I will do it. On Mon, 3 Aug 2026 at 05:46, Moritz Angermann via ghc-devs < ghc-devs@haskell.org> wrote:
Hi Teo,
Contributors should only put forth contributions that they could have created without the support of AI tooling.
I agree with that statement. Although I'd probably add reasonably to it. I do think it suffers from the same "very difficult to tell":
Is the intention that if a user does one of the things in "inappropriate use" that they're contribution would be rejected? These seem like extremely controversial suggestions to me, and also things that are very difficult to tell.
Out of curiosity, which of the inappropriate use cases listed do you find extremely controversial? I'm still operating under the premise that there is neither a right to have one's contribution reviewed, nor a right to have one's contribution included/merged. It's ultimately up to the core maintainers/folks with commit bits to decide if they deem a contribution inclusion worthy or not. Maybe this assumption is fundamentally flawed and just wrong?
Best, Moritz _______________________________________________ ghc-devs mailing list -- ghc-devs@haskell.org To unsubscribe send an email to ghc-devs-leave@haskell.org
participants (6)
-
Bryan Richter -
Hécate -
Julian Ospald -
Moritz Angermann -
Teo Camarasu -
Wolfgang Jeltsch