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

I'm sorry if this is considered too long. Please read if you intend to 
answer. Otherwise just skip to TLDR.

Am 13.05.25 um 18:53 schrieb Russ Allbery:
> Our disagreement is not about motives, at least in that sense. Our
> disagreement is over consent, and your willingness to break the social
> contract with artists and override their consent in order to get the
> supposed benefits of LLMs.
Where did you get that social contract from? I don't know about any. 
Lawmakers - elected or not - once gave copyright to society and EU 
lawmakers - now all elected - took parts of it away again. Easy like 
that. 2021. Nobody really protested. Poof, gone.

We have a social contract to not let our citizen starve (artist or not) 
and we are still sticking to it. But a social contract about the 
extension of copyright? Certainly not.

Looking at the present government in the US and its entanglemant with 
big tech I'm sure the same will happen in the US - "one way or the 
other" (insert his annoying voice here). No social contract will hold 
against big tech power - if there ever was one in the US.

Looking at the rest of the world I dare to state that most countries 
weren't very much invested in copyright anyway. East Asia especially. 
Sorry if I can't look at every country in detail.

> a sort of populism that sacrifices the economic model that artists
> currently rely on to survive in the name of empowering the masses. You
> went so far as to start talking about LLMs as magic copyright-washing
> machines that would let you destroy the basis of copyright and use other
> people's works for whatever you want. 

Every EU citizen is legally entitled to do exactly that - as far as text 
and data mining is concerned of course. Law is a political construct 
made by powerful people and can change in the blink of an eye - and it 
did. As you wrote yourself - if I remember correctly - Free Software 
licenses (and thus the DFSG) were a hack *against* copyright. So why now 
uphold it when the law is changing? Because otherwise artists in the US 
would starve? If yes, do something about it without impeding other 
peoples freedom (use taxes to sustain them like we do in the EU).

> I'm sure that what you*think* you're doing is democratizing art. People
> who hold this position always think that. I think you have been duped into
> a politics of art without the artist, a belief that one can decouple the
> creation of meaningful art from artist control of their work by turning
> artistic work into a sort of factory work-for-hire regime, and that this
> is somehow a good thing for the world. I don't think that*you personally*
> would classify your position that way; I suspect you just think of it as
> expanding the commons and giving artists better tools. But in our current
> society this is, to me, the very obvious and predictable outcome of your
> political position.

An artists work does not and should not depend on copyright. Copyright 
only empowers the already powerful because anybody else cannot afford 
litigation. Sustaining art is a decision a society has to take and to 
finance. As is done in the EU (lots to criticise but that's not the 
topic here).

> I feel some
> obligation to try to argue for an ethical position in the places where I
> may have some influence.

Deriving an ethical position from the design of a random legal artifact 
seems strange to me.

> But I certainly don't think your position is anti-human! Quite the
> contrary: I think it's techno-populist in all of the worst senses of
> populism, the sort of populism that is being used as an unwitting tool by
> the richest people on the planet who resent that art currently (unevenly,
> imperfectly) favors labor over capital, and whose goal is to ensure
> capital reigns surpreme in all areas of society.

I think this is an unfair attack. Debian is a project that primarily 
serves its users. All decisions should be taken with the users best 
interests in mind. Programs with machine-learning-derived components 
(I'm trying to avoid the AI buzzword) have lots of benefits for our 
users: for the blind, the deaf, the otherwise handicapped, for 
everybodies daily work, for new ways of creativity ... and we can help 
them by giving them tools that are NOT controlled by big capital but 
cared for by the Free Software Community. Requiring the afaik everywhere 
illegal redistribution of training data effectively sabotages that goal.

TLDR: All these political/ethical arguments are clinging to a certain 
(US) view of society and are invalid at least in Europe. EU lawmakers 
made a decision and EU developers will follow it. Debian might kick 
itself out of orbit though.

 From the DFSG only two relevant questions remain:
1. Is training data source code? I think it would be a stretch. Souce 
code and training data are different categories and not comparable. 
Nobody else qualifies data as source code. Or do they?
2. What is the preferred form of modification? This is IMHO the 
deciding, relevant question.
Aigars says weights and I've heard that from several other people active 
in machine learning. OSI says the same.
Mo Zhu says training data is. I haven't heard that from anybody else.

I'm not an AI expert so I don't know. I would like to see this discussed 
for a change.

Ilu