Comment by 99failures

So really it boils down to attribution, therefore if GitHub were to disclose attribution to all the copyright owners of the code used to train the model, then this issue will be mute.

It will be a long list, but just a list.