Renaud Allard <renaud@allard.it> wrote:

> > I think the model was mostly trained on scientific papers on this task
> > and it does not fall under the GPL either to follow instructions in
> > scientific papers I guess.
> 
> OpenBSD is the project for which ambiguous code is never
> imported. There have been precedents.
> 
> Read carefully https://www.openbsd.org/policy.html

Renaud, I think you have it right.

The document already makes it clear.

It talks about how we've built a combined system from components which
are usable and re-distributable with permissions (granted non-exclusively)
by all the various people, authors, and creators.

There are three important aspects:

      - who can provide rights to that work
      - the rights required by the project
      - who can claim Copyright on a work to provide rights

Copyright is a legal right bestowed upon a human, or a collection of
humans, which grants them exclusive control and -- by default --
prevents re-use and re-distribution.

Check the treaties and the laws, and you'll find that Copyright is all
about human output.  (There are people dancing around trying to twist the
laws, but their comments are in Azimov territory).

In our software ecosystem we require the human (recognized under the
law) to grant us a number of rights.  Actually, we require those rights
to be granted to everyone not just us (because there is no point in us
participating in the redistributing of something that someone else
cannot again redistribute).  An ISC block does the best job showing the
rights being granted.

 * Copyright (c) 2015 Nicholas Marriott <nicholas.marriott@gmail.com>
 *
 * Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.

This human, Nicholas Marriot, gives everyone the required rights for us
(or anyone else) to combine it into a complete system.  Otherwise, we
could not use this.

The indent(1) command is a C program formatter.  It takes input and
produces output.  AI systems are more similar to indent(1), than to the
human process, because when the AI produces something it is pulling it
from an unknown corpus of information it ingested, and it isn't just
pulling it from it's ass like a human would because the AI has no ass.
Copyright laws recognizes that humans can create, and has no such
recognition for machine processes (whether they are simple or complicated).

We all know that re-indenting code doesn't remove the old Copyright.
We must assume that re-AI'ing code doesn't remove the old Copyright.

At present, the software community and the legal community are unwilling
to accept that the product of a (commercial, hah) AI system produces is
Copyrightable by the person who merely directed the AI. 

And the AI, or AI companies, are not recognized as being able to do this
under Copyright treaties or laws, either.  Even before we get to the point
that the AI's are corpus-blenders and Copyright-blenders.

So as of today, the Copyright system does not have a way for the output
of a non-human produced set of files to contain the grant of permissions
which the OpenBSD project needs to perform combination and
redistribution.