Re: GR -- Allow AI-Assisted Contributions

Pierre-Elliott Bécue <[email protected]> Thu, 19 Feb 2026 11:58:30 +0100
Newsgroups	gmane.linux.debian.devel.vote
Message-ID	<[email protected]>
Lucas Nussbaum <[email protected]> wrote on 19/02/2026 at 10:14:54+0100:

> Hi Harlan,
>
> On 18/02/26 at 17:48 -0600, Harlan Lieberman-Berg wrote:
>> On Wed, Feb 18, 2026 at 4:38 PM Lucas Nussbaum <[email protected]> wrote:
>> 
>> > Based on recent discussions, it sounds like we probably need a
>> > discussion in the framework of a GR to understand where we stand
>> > regarding AI-assisted contributions to Debian.
>> 
>> 
>> Hello Lucas, all,
>> 
>> With all due respect to your work on drafting this, Lucas, I strongly urge
>> us not to go through with a GR on this topic yet.  This is something which
>> both us and the larger free software community is still actively debating,
>> and a GR is a tool for when a decision is ready to be made.  Having a GR on
>> this matter now is premature and will be extremely divisive -- even
>> schismatic.
>> 
>> As someone with personally strong feelings about AI, I still believe that
>> we are better off with the current status quo where we can still discuss,
>> rather than vote and decide.
>
> I disagree.
>
> First, several major organizations and projects have established policies
> or guidelines about AI-assisted contributions. Some examples:
> https://www.apache.org/legal/generative-tooling.html
> https://www.linuxfoundation.org/legal/generative-ai
> https://www.eclipse.org/projects/handbook/#genai
> https://firefox-source-docs.mozilla.org/contributing/ai-coding.html
> https://openinfra.org/legal/ai-policy/
> https://docs.fedoraproject.org/en-US/council/policy/ai-contribution-policy/
> https://discourse.llvm.org/t/rfc-llvm-ai-tool-policy-human-in-the-loop/89159
>
> Second, the current status quo within Debian is that there are many
> people who experimented with AI tools, concluded that they could help
> them with their Debian work somehow, are probably already using them for
> Debian work. But they prefer to stay mostly silent in the discussions
> and only express support in private because there's a group of people
> that vehemently rejects AI and attacks people that admit using it.
>
> What I'm seeking with this GR proposal is a middle ground that allows us
> to respectfully move forward, accommodating both the people who want to
> use AI tools in the context of Debian and share their experiences, and
> the people who want to limit their exposure to AI.
>
> I do not want anyone to quit Debian as a result of this discussion, so I
> tried to write down a proposal that goes quite far into accommodating
> both groups, and we could probably go further or have other options in
> both directions.
>
> But, as I wrote elsewhere, I view the current discussions about AI like
> discussions about politics (or VI vs Emacs, or programming languages).
> People approach these topics with vastly different backgrounds; there is
> little hope of fundamentally changing someone's core opinion.
> Consequently, while these debates are interesting, they are also
> exhausting.
>
> I believe we should focus on what is desirable specifically for the
> Debian project. We should agree to disagree on AI in general, and
> instead focus on how we can work together despite that disagreement.
>
> I don't want to rush things, but I also don't think that it's useful to
> spend a lot of time rehashing arguments.

Starting a GR thread this quick doesn't give the impression that
discussions will be able to occur peacefully and as slowly as Debian
might need.

I'm a bit unsettled by the fact that you did go this way, especially
because it doesn't give a sufficient amount of time to non-Debian
Members to give us their opinion and feedback.

-- 
PEB
signature.asc (application/pgp-signature, 853 B)