[License-discuss] Curious about derived works and AI...

[Shuji Sado]

Unlike common law countries such as the UK and the US, Japan adopts a
statutory law system.
Therefore, Japan needed to amend its Copyright Act to promptly address
copyright issues as machine learning technology became practical.
I understand that the current Japanese Copyright Act aims to comply
with the three-step test established by the Berne Convention and the
WIPO Copyright Treaty, protecting creators' rights while ensuring that
AI system developers are not discouraged from using copyrighted works.

As a result of the 2018 amendment to the Copyright Act, the act of
training AI systems and the act of generating content are judged
separately for infringement.
The act of an AI system learning from various copyrighted works is
generally considered legal under Article 30-4 of the Copyright Act,
but if a trained AI system outputs content similar to a copyrighted
work and that output is used, it is illegal. For example, if an AI
system trained on GPL code generates code similar to that GPL code,
using that similar code under a license other than the GPL would
obviously be a license violation.

Yes, there are pros and cons to allowing AI systems to learn from
copyrighted works without permission. I am also deeply troubled by
this issue.
However, when a time comes when everyone naturally possesses an AI
system they can control, and if asked, "What is the difference between
an AI learning and a human learning using their brain?" I can only
answer, "In the framework of copyright, there is no difference."

Perens-san,
I listened to your Post-Open lecture at the OpenUK event. It was an
interesting talk. However, I also thought it is a very long journey.
At least, it is not a problem that can be solved by the mechanism of
licensing(not contract) alone.
I understood that the focus is on the issue of the imbalance of
interests between Big Tech and small communities, but it also seemed
like a problem that cannot be solved by any single technology or idea.
Personally, I believe we have to steadily promote a free and fair
system in every aspect, step by step.

--
Shuji Sado
Chairman, Open Source Group Japan
https://opensource.jp/
https://shujisado.com/

2024/06/01 1:12 Bruce Perens via License-discuss
<license-discuss at lists.opensource.org>:
> On Fri, May 31, 2024 at 5:08 AM Stefano Maffulli <stefano at maffulli.net> wrote:
>>
>> You should realize that you're implicitly arguing for an extension of
>> copyright at the expense of freedom of research, open knowledge and open
>> science, before Open Source.
>
> I'm not convinced. I think we are talking about existing implications of copyright that should be resolved through case law in the United States. Obviously the cases are in progress, so that's not an entirely satisfying answer today. I also think it's inevitable that WIPO will get involved and that sometime in the future Japan may have to make a decision based on their remaining within treaties and retaining the trade options that come with that - which are a pretty big incentive. I'm not saying I approve of the WIPO regime, I just think that's how things will work out.
>
>> But you need to realize that  the mega corps already have
>> swallowed your code and writings and pictures and use patters, graph of
>> friends, etc, through perfectly legal means, like the GitHub terms of
>> service. And they'll continue to do so with bilateral deals (Reddit,
>> Stackoverflow, etc.)
>
>
> This is a really good point and one that we have the full power to address, and to promote the solution rather than just complain: we should not heedlessly enter into corporate terms of service just to do our operations. And that's what we're doing now, it's heedless, most of us have never even read those terms. What alternative site do readers recommend? I also need an online discussion / mailing list forum without such terms.
>
> So, why can't you do that, Stef? I think one of your "mega corp" sponsors might be offended if you did. I did run into such a constraint while working on the ETSI issue for OSI. People need to be aware that there are places that corporate-sponsored Open Source non-profits can't tread, and leadership on those issues must come from elsewhere.
>
>>
>> Your argument is going to deprive open research groups like Eleuther AI,
>> LLM360, LAION and future ones of an opportunity to compete with the mega
>> corps, imposing restrictions to text and data mining that will limit
>> only the **real** Open Source AI.
>
>
> In the draft Post-Open Zero Cost License  ( https://perens.com/static/DEVELOPMENT_LICENSE.txt?v=2024_05_24.1 ) I made it so that a machine learning model could be trained if the model was itself under the same license. See 2.6, 2.7, 2.9, 2.12(c), 2.13, 2.14, 3.3(f), 4.8, 4.14. That license, of course, is a response to the corporate abuses you are complaining about.
>
> I will pursue the documents you recommended.
>
>     Thanks
>
>     Bruce
> _______________________________________________
> The opinions expressed in this email are those of the sender and not necessarily those of the Open Source Initiative. Official statements by the Open Source Initiative will be sent from an opensource.org email address.
>
> License-discuss mailing list
> License-discuss at lists.opensource.org
> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org