Interesting to see the mad scramble of proprietary software / silo companies updating their ToSes to say you explicitly consent to your data being used to train #AI -- while also taking the approach that they don't need consent (or compliance) to train on copyleft licensed works.