It's either intentional, which would be puzzling and unsettling, or it's a bug which has gone unnoticed. In any case it is proof that big tech is in no shape to take on the responsibility for moderating discourse on the internet. This reminds me of the bug that falls into a typewriter in the beginning of the movie "Brazil" which causes a spelling error and the arrest and execution of a random innocent person. Granted, this type of automated banning without any ability to involve a real human is not costing any lives (yet), but I am increasingly worried about how big tech is becoming a Kafkaesque lawnmower. One thing is to deliberately censor speech that you do not like, another is to design a system where innocent and important speech is silently censored and noone in charge even notices.