Re: [paho-dev] AI contributions - policy needed?

[Pau Figueras <pau.figueras@xxxxxxxxxxxxxx>]

Hi,

I think we should update the bug report templates and add the AI Contributions part to the CONTRIBUTING.md file as Matt suggested.

I'll go ahead and create an issue on the repo to do just that.

It's not like the LLMs are going to read that, but at least we'll have strong grounds to reject some PR's and issues.

Pau Figueras | Software Engineer

---

De: paho-dev <paho-dev-bounces@xxxxxxxxxxx> de part de Matt Brittan via paho-dev <paho-dev@xxxxxxxxxxx>
Enviat el: dilluns, 21 de setembre de 2026 3:28
Per a: General development discussions for paho project <paho-dev@xxxxxxxxxxx>
A/c: Matt Brittan <matt@xxxxxxxxxx>
Tema: Re: [paho-dev] AI contributions - policy needed?

 

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi All,

Personally, I'm not convinced AI generated submissions can possibly be allowed under the Eclipse CLA:

You will only Submit Contributions where You have authored 100% of the content

It looks like the CLA has been updated since then. That clause is in V1.0.0 (https://www.eclipse.org/legal/eca/cla.html), but not in the current version 3.2.0  (https://www.eclipse.org/legal/eca/ - I believe the CLA became the ECA in 2018). The later version's "Developer Certificate of Origin" includes:

The contribution was created in whole or in part by me and I have the right to submit it under the open source license indicated in the file; or ...

This appears to support AI use (I still have concerns regarding training data rights). Eclipse offers GitHub Copilot Enterprise for Eclipse Committers which can review change requests (and make improvements); therefore, the Foundation is actively promoting AI use.

I'm seeing a few issues in the Paho Go repos. One user raised 28 AI generated issues within two hours (these can be condensed into a couple of issues, but condensing them will take me a while). I'm sure the Python library is being hit much harder!

Whilst I'm concerned about AI's impact on the project, I don't think we can, or should try to, avoid it. Suggestions:

• On the Go repos I've already updated CONTRIBUTING.md to include: "AI Contributions - Please see the Contributing Content Generated by an AI section of the Eclipse handbook for guidance. Take special note of the disclosure guidelines."
• Update the Issue and PR templates to provide direction to agents (e.g. I added "Please do not raise a PR if you are not prepared to sign the ECA (we will be unable to accept it)." in the hope that this will prompt agents to alert the contributor of the need to sign the agreement).
• Maybe something along the lines of Greg's suggestion: i.e. "Was AI used in the creation of this issue/PR? If so, please provide details."

Matt

On Mon, 21 Sept 2026 at 10:58, Greg Troxel via paho-dev <paho-dev@xxxxxxxxxxx> wrote:

Frédéric Desbiens <frederic.desbiens@xxxxxxxxxxxxxxxxxxxxxx> writes:

> Hi everyone.
>
> The Eclipse project handbook contains generative AI guidelines for
> committers <https://www.eclipse.org/projects/handbook/#genai>. If AI

Wow.  That reads as LLM-generated pro-LLM propaganda.  I quote briefly:

  The Eclipse Foundation values the human-centric approach to creativity,
  where Artificial Intelligence (AI) serves as a tool to enhance
  efficiency. By embracing the power of AI, we prioritise the empowerment
  of individuals to fully explore and express their creative
  potential. We believe in leveraging AI to optimise workflows,
  streamline processes, and minimise mundane tasks, allowing creators to
  focus on what truly matters: their imaginative pursuits.

There's no discussion of respecting the humans that will perform review,
by making the contribution as concise as if it were fully
human-generated.

> contributions are accepted, then they should comply with the rules
> highlighted in the handbook. The short version is: any usage of AI should
> be properly disclosed, and the human contributor remains responsible.

Somehow the CLA that says the offered content was wholly created by the
author has vanished.

> Naturally, a project could decide to exclude AI contributions. In that
> case, the project team should take a proper vote on this matter and make
> this policy known on the website, on GitHub, and through other channels as
> well.

In another context, I saw an observation that this "can exclude LLM
content" framing is already begging the question, vs prohibited unless
there is a policy to accept it.  I don't expect anyone to take that
comment seriously, given the staggeringly pro-LLM text in the handbook.
_______________________________________________
paho-dev mailing list
paho-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/paho-dev