Hello!

It has been very nice to read the thoughts and opinions, and I have
comments because I have not found anything in the ASF guidelines or related
documents regarding the watermark that is being required under the AI Act’s
Transparency Code. Different regulators around the world are increasingly
mandating that such watermarks be embedded into AI-generated code.

>From my perspective, it is clear that because the ASF is a U.S.-based
entity, it must comply with the laws, regulations, and policies of its host
country. However, this raises an important question about code contributed
from foreign jurisdictions and how such contributions should be handled in
light of local regulations.

I believe the guidelines, whether from the ASF or the Apache Fineract
project, should explicitly address AI usage. These guidelines should also
be treated as living documents rather than static ones, given how rapidly
the AI landscape is evolving. Every user, contributor, reviewer, and
adopter should be aware that different jurisdictions have varying
approaches to AI generation and usage, particularly with respect to
intellectual property, copyright, and transparency requirements.

I would also be interested to know whether the ASF infrastructure team has
any recommendations regarding GitHub Actions, tools, or other mechanisms
that could help identify AI-generated code (especially in light of the
emerging watermarking requirements), and a final thought... so then the
Apache Fineract will have the ticket in this regulated jurisdiction "Apache
Fineract contains AI generated code and can make mistakes. Please check the
cited sources again" ?

Regards,
Victor Romero

El mar, 1 sept 2026 a las 16:33, Adam Monsen (<[email protected]>) escribió:

> James Dailey wrote:
>
>> https://rai.apache.org/ai-generated-code.html (again DRAFT)
>>
>
>  +1. This is excellent.
>

Reply via email to