So the word/phrase filtering in Pleroma is how content filtering ought to work. That is, the user decides if they want to hide certain posts and not the author of the post trying to imagine what all the users might think of it and adding CWs accordingly.