The same has happened to me many times. Don't take it personally. AI is not the problem. Use of AI, in fact, would be a big improvement. The iTalki algorithm is far less sophisticated than that. They have a list (that they will not reveal) of banned words. I learned a bit about it by searching online at websites like reddit where people are free to discuss such things. Some of the words on the list are surprising. For example, there is a very common man's name that also refers to a part of the body. Use it and you will be banned. Avoid words that could possibly be used in some sort of threat and words that could be used to describe a weapon or bodily harm. Avoid sorrow and suffering. That eliminates a lot and sometimes makes it difficult to say what you want to say. Use roundabout circumlocutions to avoid words that algorithm might not like. For example, there is a word that starts with "d" and ends in "h" that refers to the end of a life. I think that word is no longer banned but it once was.

I've experienced what you describe. The only advice I can give you is that you should always be aware that the moderation is automated here and the filters are extremely sensitive. If you have any doubts about how the moderation software might react to something you post here, either don't post it or censor the words that might trigger moderation by substituting symbols in the place of some the letters.

My guess would be that the AI missfired.

One other thing - if you include a link it often seems to get the post banned.

There is AI but sometimes it fails. I've been banned for a whole week because the AI incorrectly flagged my comment. Send a message to italki support. They will go over your comment and reverse the ban if there is genuinely nothing wrong with it. If they believe you violated a community guideline, they will tell you what rule you broke.