creators of LLMs would like to know [...] that they haven’t been trained on copyrighted data.
I'm not quite sure about that LOL
Hasn't Google recently announced that the whole internet now belongs to them for the purpose of training their next models?