@Jerry @alainamartin Good point, I completely forgot about this recent news.
However, as you say, you can never be certain that an instance is using cloudflare or if they are using it, can't be sure that they're using this new feature.
Alaina, to your question, we just don't know if a company is already using Fediverse data to train AI.
There's lots of folks, including myself, that try to keep track of such things, but due to the vastness of Fedi, it'll never be a sure thing.