Re: Proposal -- Interpretation of DFSG on Artificial Intelligence (AI) Models

Matthias Urlichs <[email protected]> Mon, 28 Apr 2025 02:49:14 +0200

Newsgroups	gmane.linux.debian.devel.vote
Message-ID	<[email protected]>

On 28.04.25 00:02, Russ Allbery wrote:
> Some of
> them, however, will be trained on things that are widely recognized to be
> non-copyrightable facts, such as records of backgammon, chess, or go
> games.
However, the manifested collection of such games might be covered by 
some database aggregate copyright-or-something-like-it law, which 
doesn't help.
> However, even that is tricky, because the*annotations* on chess games can
> be copyrighted.

… which is only a problem if the chess engine considers them for 
training, which I assume they don't, so we could conceivable DFSG-ize 
the data.

The fact remains that our builders will be unable to reproduce the 
resulting network, for well-known practical reasons. Thus we 
mostly-have-to-trust the original publisher that their network has been 
built as documented (or even "documented" given the status of gnubg). In 
practice this is not a problem for a Backgammon engine, or even for 
Tesseract because any serious use case supports, if not requires, human 
verification of the result — but how sure can I be that a LLM intended 
for home automation doesn't contain an Open Sesame backdoor that unlocks 
my *home*'s back door?

-- 
-- regards
-- 
-- Matthias Urlichs

matthias.vcf (text/vcard, 195 B)

BEGIN:VCARD
VERSION:4.0
N:Urlichs;Matthias;;;
NICKNAME:Smurf
EMAIL;PREF=1:[email protected]
TEL;TYPE=work;VALUE=TEXT:+49 911 59818 0
URL;TYPE=home:https://matthias.urlichs.de
END:VCARD

OpenPGP_signature.asc (application/pgp-signature, 840 B)