Nathan, I think that's a good idea. The licensing aspect of AI-generated code is a huge gray area, even in other fields. I'm pretty sure copyright disputes about it are still being debated all over the world. It's best to err on the side of caution and request the contributions be substantially human-written, I agree. We can add those points to the Matplotlib policy if we end up integrating this into our guidelines.
Thanks for your human-generated input :) Matteo On Sat, Feb 21, 2026 at 11:32 AM Nathan Hartman <[email protected]> wrote: > +1, yes, this is a good start, with Tomek's suggested text. It needs a > little bit more, though: > > AI has been shown to regurgitate someone's code verbatim. Now, if it > regurgitates Apache-2.0 licensed code, that might be OK, but what if it > regurgitates code with other licenses such as GPL, or some proprietary > "code available" license? The ASF's rules require that contributors' > submissions be their own. I think we need to emphasize that. > > Tomek suggests writing that the contributor should understand and be able > to explain the change. Matplotlib says the contributor should add value > with their own competency. We should perhaps emphasize that contributors > can use AI to aid their work but that their submission must be > *substantially* theirs. > > In other words: > > * using AI for brainstorming, prototyping, finding information/examples, as > a typing assistant--anything where the submission is the result of the > person's own intellect--is OK. > > * using AI to generate a PR and then making some minor edit to pretend it's > your own, is NOT ok. > > P.S., this message was written entirely by me, WITHOUT any AI. Yeah, I'm > old school like that. :-) > > Cheers, > Nathan > > On Sat, Feb 21, 2026 at 7:47 AM Alin Jerpelea <[email protected]> wrote: > > > +1 > > > > a good start > > > > Brst regards > > Alin > > > > On Sat, 21 Feb 2026, 08:24 Tomek CEDRO, <[email protected]> wrote: > > > > > +1 :-) > > > > > > Looks like we have a good reference point, and we may include this > > > part to NuttX Contributing Guide? > > > > > > """ > > > Restrictions on Generative AI Usage > > > > > > We expect authentic engagement in our community. > > > > > > Do not post output from Large Language Models or similar generative AI > > > as comments on GitHub or our discourse server, as such comments tend > > > to be formulaic and low content. > > > > > > If you use generative AI tools as an aid in developing code or > > > documentation changes, ensure that you fully understand the proposed > > > changes and can explain why they are the correct approach. > > > > > > Make sure you have added value based on your personal competency to > > > your contributions. Just taking some input, feeding it to an AI and > > > posting the result is not of value to the project. To preserve > > > precious core developer capacity, we reserve the right to rigorously > > > reject seemingly AI generated low-value contributions. > > > > > > In particular, it is also strictly forbidden to post AI generated > > > content to issues or PRs via automated tooling such as bots or agents. > > > We may ban such users and/or report them to GitHub. > > > """ > > > > > > What I also really really like in that Matplotlib's Contributing Guide > > > is the "difficulty" Issues label! We may mark tasks/todos as the > > > github issues, as we do currently, but assign "difficulty" label for > > > them, so we could redirect newcomers to solve "easy" issues in the > > > first place to see how they work. This should eliminate completely new > > > people taking complex tasks without understating the basics and > > > potentially breaking stuff as we see recently. This would also > > > eliminate new people showing up with invented long term tasks just to > > > get GSoC but knowing noting about NuttX and having no prior > > > experience, etc. This would help onboard new people, get them familiar > > > with the project internals, contributing process, and we could have > > > tasks solved step by step :-) > > > > > > -- > > > CeDeROM, SQ7MHZ, http://www.tomek.cedro.info > > > > > > On Sat, Feb 21, 2026 at 5:03 AM Matteo Golin <[email protected]> > > > wrote: > > > > > > > > Separate from the other thread: > > > > https://www.mail-archive.com/[email protected]/msg14332.html > > > > And starting fresh from the previous discussion: > > > > https://www.mail-archive.com/[email protected]/msg14173.html > > > > > > > > I would like to propose the adoption of Matplotlib's AI policy [1] on > > > > NuttX. I believe it is a pretty reasonable middle-ground where: > > > > - Completely AI-generated PRs (code/descriptions) and comments via > > > > automated tooling, or in other words, slop, is not allowed on any > > > > repository/official communication channel of the project > > > > - Use of generative AI to aid coding, documentation writing, etc. is > > > > allowed but comes with the caveat that the contributor must > understand > > > > their change and should have added some value with their personal > > > competency > > > > - Reserves the right to reject low-effort, AI-generated contributions > > > > > > > > [1]: > > > > > > > > > > https://matplotlib.org/devdocs/devel/contribute.html#restrictions-on-generative-ai-usage > > > > > > > > Let me know what you think, > > > > Matteo > > > > > >
