@james 👆This is the whole problem with AI generated anything right now. It is a giant legal grey area. Unless you know that the training data came from purely public domain works, you can never know if you are violating the original license of something someone else made. This is the biggest problem with GitHub copilot right now. It is trained on code with all sorts of different licenses, and as a result, it effectively serves as a tool for separating code from licenses.