[epistasis]

In this thread we are talking about the US military, China, North Korea, etc. using the models, and whether anyone should give a shit about ToS.

I'm not aware of AI companies violating any software ToS or the GPL, but would like to know more.

(And I think that everybody should follow the GPL on software that is licensed that way, just as they should the ToS.)

[jowea]

There's been a big controversy over training AI on copyleft code hosted on places like GitHub and generating code which could be interpreted as a derivative work.

[epistasis]

I have not heard discussion of that, but trying to reconstruct what it would be my guess is: AI companies with closed, non-distributed models in the clear with the GPL (but not AGPL code). But Meta would not be able to distribute their weights without also distributing the GPL code with it that went into the training? Is the controversy that it "linked" all the other source code that the model was also trained with and therefore all thus source code must be distributed too if the model weights are distributed?

I'm really curious because my best guesses sound ridiculous to me when I type them out.

[jowea]

Random article about the topic: https://fossa.com/blog/generative-ai-and-software-developmen...

> All the while, there’s been debate about whether engineering teams that use generative AI are at risk of violating open source software licensing requirements. (Generative AI tools train, in part, on open source libraries, which leads to questions about whether generative AI output is subject to the same licenses as the training material.)