[Biopython] Generative AI policy for contributions to Biopython

[Peter Cock]

Exactly - Andew noted he has seen real examples of this problem:

https://mailman.open-bio.org/pipermail/biopython/2026-April/017116.html

In both his examples a naive AI using contributor might presume there is no
copyright claim (and check the box on our pull request template), where a
more informed opinion raised serious concerns.

Even the new Linux plan to ask the contributor to take full responsibility for
the contribution and sign-off on their Developer Certificate of Origin (DCO)
has this problem - going back to the start of this thread:

https://mailman.open-bio.org/pipermail/biopython/2026-April/017113.html

These risks existed before generative AI but will be much more of an issue now.

Peter

On Mon, Apr 27, 2026 at 3:19 PM Hilmar Lapp <hlapp at drycafe.net> wrote:
>
> Just as a note, what is needed is not just an assertion of being free of license restrictions, but also being free of someone else’s (who isn’t the contributor) copyright claim, because Biopython’s license declaration asserts copyright on behalf of “The Biopython Contributors" for all code.
>
> That being said, the last sentence below holds regardless, namely that this isn’t a new requirement since using AI has become an option.
>
> The part that arguably is, however, new is that if I personally write some piece of code, I do know and thus can sincerely assert whether and to what extent I wrote it copying someone else’s code, and whose code I took inspiration from. LLMs _can_ generate code that is new, but memorization from learning is a known problem and I’m not sure how a contributor using AI would be able to sincerely assert that some part of their code isn’t already copyrighted by someone else, even if there are no license restrictions.
>
> It seems to me that this issue is usually overlooked. If copyright is held by the individual contributor, as I think the Biopython copyright declaration means to say, then this is probably a non-issue at least in the legal sense, meaning that no copyright violation liability rests with the project as a whole.
>
> Just to be clear, IANAL!
>
>   -hilmar
>
> On Sat, Apr 25, 2026 at 12:44 AM Michiel de Hoon <mjldehoon at yahoo.com> wrote:
>
>
> - The PR submitter takes responsibility for guaranteeing that the contributed code is free of license restrictions, and therefore can be released under Biopython's license. But the same requirement holds for any code contributed to Biopython, not just AI-generated code.
>
>
> --
> Hilmar Lapp -:- lappland.io
>
>
>