Re: Ballot option: Allow AI-Assisted Contributions

The Wanderer <[email protected]> Fri, 24 Jul 2026 11:33:00 -0400
Newsgroups	gmane.linux.debian.devel.vote
Organization	This space intentionally left blank.
Message-ID	<[email protected]>
On 2026-07-24 at 05:03, Stefano Zacchiroli wrote:

> On Fri, Jul 24, 2026 at 10:45:15AM +0900, Simon Richter wrote:
> 
>> if the decision ends up being that it cannot [be copyrighted], we
>> need to delete the corresponding copyright attributions (as far as
>> they don't overlap with valid ones).
> 
> This is not true, on at least a couple of levels.
> 
> First, even if courts will confirm that LLM output cannot be
> copyrighted (in the US…), that doesn't mean that a contribution
> labeled with "Assisted-by:" is not copyrightable.  For example, the
> submitted might have reworked the contribution after generation,
> enough for the contribution to be copyrightable.
> 
> Second, "not copyrightable" doesn't equate "not DSFG free". For 
> instance, public domain data is not copyrightable, but we can (and
> do) have and distribute them in Debian. (And: no, you don't need sui
> generis database rights to be able to do this.)

My understanding - which could be incorrect - is that the original
motivation/justification for "Assisted-by:" tags in the context where I
understand them to have been first been used, that being the Linux
kernel, was not about "in case the content is ruled not copyrighted",
but "in case the content is ruled to be copyrighted by virtue of the LLM
that produced it having been trained on copyrighted material". That is,
the idea was "if courts rule that the output of LLMs trained on
copyrighted material is subject to the copyrights and licenses of that
original material, we'll have a legally-enforceable obligation to be
able to identify which things in our source trees are derived from the
output of those LLMs, and remove that tainted content".

Whether or not that's a realistic concern (then or now, since IIRC there
may have been some rulings which could have gone that way but didn't?)
is a different matter, but that at least seems like a distinctly
different consideration from "can it be copyrighted at all?".

> It certainly might happen that future court decisions will force us
> to remove content from Debian, but there is nothing new here: it
> might happen also with content that has been produced with no LLM
> help at all.

The above is about a channel of vulnerability to such need for removal
that probably only exists (and almost certainly could only be annotated
in this manner) with LLMs, however.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man.         -- George Bernard Shaw
signature.asc (application/pgp-signature, 833 B)