Re: GR -- Allow AI-Assisted Contributions (was Re: DFSG-new-queue is AI-generated)

! Responses by people wanting to defend, promote, etc. so-called “AI”
! are *explicitly* neither solicitated nor welcome (last time, aigarius
! managed to derail the discussion about the counter-proposal), as they
! have a different proposal to express themselves.


On Thu, 26 Feb 2026, […] wrote:

> There's a discussion on debian-vote (at least) on AI policy in
> Debian, which is a much better venue.

Thanks for noting.

I’d like to resubmit my counter-proposal from almost a year ago.
As it seems, it’s even more “counter” this time.

The last version thereof was:
https://lists.debian.org/debian-vote/2025/05/msg00145.html

There was a thread of discussion related to that. Not sure whether
we need to pick up anything more from that. (There was a response
from soren about the whole copyright thing; we know more nowadays:
training a model for analytic use (to discover trends, etc.) is
exempt under the TDM exception to copyright, unless explicitly
opted-out in a machine-readable way (which didn’t exist back then
and which likely are still not honoured and whereof there are at
least three, these days…); using a model to produce output isn’t
covered by TDM, though, so it’d need an explicit licence or use PD
works only; the US link from… this thread somewhere, IIRC… also
doesn’t give such permission).

zigo wanted to be more permissive for non-free but otherwise
agreed-ish. Perhaps we can work on this together and then either
come to some agreement or split it up into two ballot options
that only differ in handling of non-free.

I don’t have much time or spoons for this, unfortunately.

bye,
//mirabilos
-- 
Sometimes they [people] care too much: pretty printers [and syntax highligh-
ting, d.A.] mechanically produce pretty output that accentuates irrelevant
detail in the program, which is as sensible as putting all the prepositions
in English text in bold font.	-- Rob Pike in "Notes on Programming in C"