What about all the training data that's no licensed for commercial use. It's probably quite likely that they scanned all of github's public repositories. Includ

What about all the training data that's no licensed for commercial use. It's probably quite likely that they scanned all of github's public repositories. Including many projects that don't give permission to use commercially. Surely that's legally problematic?!