[QGIS-Developer] Tightening AI submission policy?

[Greg Troxel]

Nyall Dawson via QGIS-Developer <qgis-developer at lists.osgeo.org> writes:

> On Tue, 23 Jun 2026 at 19:06, Julien Cabieces <julien.cabieces at oslandia.com>
> wrote:
>>
>>
>> Hi all,
>>
>> +1 as well
>>
>> > - LLMs may only be used as an improved auto-completion mechanism, or
>> > for repeated tasks (mechanical refactoring) that could potentially be
>> > completed with a deterministic algorithm.
>>
>> Shall we not also allow AI when it's used as a chatbot to answer
>> specific/technical questions.
>
> I think we should, but maybe we don't need to explicitly state that? It's
> unfortunate, but I've personally found that the web in 2026 is a wasteland
> and search engines are basically useless for finding out information
> anymore. Where once we could turn to stackexchange or blogs (or even
> upstream documentation) there's just nothing out there anymore. So
> realistically, I don't think we should block this. 😭

I think the issue here is the scope of policy.  "may only be used as",
as I read it, obviously is about LLM-generated text that makes its way
to a project forum.  I do not read it as affecting what people do before
they decide to submit something.

If someone engages with an LLM, and then submits a PR or email *that
does not contain LLM-generated text*, then I don't think qgis as a
project should, today, ban that.

If someone posts a question to a list, and somebody else feeds that into
an LLM and posts the response in an email, that should absolutely be
prohibited.

It seems easy to adjust the text, which was originally in the context of
PR content.