Since all these AI bots are trying to get data for training out of these websites, is it worth putting in some hidden HTML tags with random data to poison the training set? I'm not sure how big of an impact it would have, or if the bots would even see it, but it might be interesting to try.
You could have a prompts.txt and responses.txt and just pick random ones for each on every page load, and just hide them somewhere.
At the very least, it might annoy the data collectors a bit?