Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.project > #14231 > unrolled thread

Re: [Discussion] AI-generated content within Debian

Started byGerardo Ballabio <gerardo.ballabio@gmail.com>
First post2026-02-19 12:50 +0100
Last post2026-02-19 18:30 +0100
Articles 2 — 2 participants

Back to article view | Back to linux.debian.project


Contents

  Re: [Discussion] AI-generated content within Debian Gerardo Ballabio <gerardo.ballabio@gmail.com> - 2026-02-19 12:50 +0100
    Re: [Discussion] AI-generated content within Debian <tomas@tuxteam.de> - 2026-02-19 18:30 +0100

#14231 — Re: [Discussion] AI-generated content within Debian

FromGerardo Ballabio <gerardo.ballabio@gmail.com>
Date2026-02-19 12:50 +0100
SubjectRe: [Discussion] AI-generated content within Debian
Message-ID<MqaHD-1KAw-5@gated-at.bofh.it>
Replying only to a specific point:

Pierre-Elliot Bécue wrote:
> the way
> some FOSS licenses work don't allow for bits of the software to be
> unlicensed.

> Let's take GPL's example. GPL is what some external people call as
> "contaminating". Essentially, if one wants to add AI-generated
> contribution to a GPL-licensed software, then these additions must also
> be licensed under GPL, which is not possible in the U.S!

I don't get this.
As I understand, it has always been considered legitimate to combine
GPL software with more liberally licensed software (e.g., BSD) and
release the combination under the GPL. Thus, the more liberally
licensed part becomes effectively "dual-licensed" -- you can use it
under the GPL, with or without the rest of the combination, or you can
strip it from the combination and use it under the original license.
I don't see this as a different case. In practice, non-copyrighted
software is "licensed under the ultimate liberal license": since
there's no copyright owner who can restrict what you can do with it,
you can do *anything*.
I'm pretty sure there are already instances of public-domain code
(e.g., code written by employees of the US government) incorporated
into GPL projects. That has never been a problem.

(By the way I wouldn't call this "unlicensed" since that means quite
the opposite: software *with* a valid copyright and without a license
cannot be used at all.)

Gerardo

[toc] | [next] | [standalone]


#14233

From<tomas@tuxteam.de>
Date2026-02-19 18:30 +0100
Message-ID<Mqg0F-1Ogw-3@gated-at.bofh.it>
In reply to#14231

[Multipart message — attachments visible in raw view] — view raw

On Thu, Feb 19, 2026 at 12:33:40PM +0100, Gerardo Ballabio wrote:
> Replying only to a specific point:
> 
> Pierre-Elliot Bécue wrote:
> > the way
> > some FOSS licenses work don't allow for bits of the software to be
> > unlicensed.

Wait: "unlicensed", by default means that nobody has a license
to redistribute the work (pretty much the opposite to, say, CC0).

> > Let's take GPL's example. GPL is what some external people call as
> > "contaminating". Essentially, if one wants to add AI-generated
> > contribution to a GPL-licensed software, then these additions must also
> > be licensed under GPL, which is not possible in the U.S!

Not "the additions", but the "combined work". The additions stay under
whatever license they are under.

> I don't get this.
> As I understand, it has always been considered legitimate to combine
> GPL software with more liberally licensed software (e.g., BSD) and
> release the combination under the GPL.

Exactly.

>                                        Thus, the more liberally
> licensed part [...]

Can we stop calling that "more liberal"? MIT et al give more rights
to the distributor, but allow them to take rights away from the user,
so "liberal" here has clearly a slant in favour of the distributor.

[agree with the rest]

Cheers
-- 
t

[toc] | [prev] | [standalone]


Back to top | Article view | linux.debian.project


csiph-web