Ask HN: FOSS licenses which disallow inclusion in Machine Learning datasets?

RobotToaster · on Nov 7, 2022

Not so much disallows, but it reminds me of a specific additional clause in the reprap license[1]:

"If any part of RepRap covered by the GPL is used to train any AI, then all the products of that AI are derivative works of RepRap and must comply with the parts of the GPL on derivative works."

[1]: https://reprap.org/wiki/RepRapGPLLicence

mytailorisrich · on Nov 7, 2022

The question is then: what is this clause worth?

I don't think that using code to train an AI with results in any code produced by that AI to somehow carry the license of that training code.

If I read the Linux source code to learn C does that mean that any C code I then write is GPL'd? No.

Now of course does acceptance of that license a contractual agreement to treat AI-produced code as derivative work?

These are all questions for lawyers...

speedgoose · on Nov 7, 2022

This license is confusing as it contradicts itself. They say it’s GPL and then add clauses making it not GPL while still saying it’s GPL.

I wouldn’t touch any software written with such a license but if I must, it will say it’s vanilla GPL and ignore the confusing and very extreme second license:

> RepRap is free; you can redistribute it and/or modify it under the terms of the GNU General Public Licence as published by the Free Software Foundation; either Version 2 of the Licence, or (at your option) any later version.

> The GNU General Public Licence (GPL) under which RepRap is licenced…

pabs3 · on Nov 8, 2022

Such a license would not meet the Open Source Definition, in particular items 5 and 6:

https://opensource.org/osd