When taken in context it looks like an innocuous tweet but to a twitter censorship algorithm that isn't aware of the context it looks like a command for someone to die.
Seconded. It happened to me a while ago; I said in French (in a reply to something): "que les antivax meurent n'est pas étonnant", which means "that the antivaxxers would die isn't surprising" but which was apparently interpreted by Twitter as a death wish on the antivaxxers (it could indeed sound like one if you ignored the last part of the sentence).
The tweet was immediately flagged, I believe by an automatic Twitter system; I learned that it's best to not use the word "die" in any language and any context, because the "AI" systems are incredibly limited and stupid, and they flag anything that may be construed in a negative way.
Twitter then asked me to delete the tweet, but what it didn't tell me is that this would be an admission of guilt. I got suspended for a number of days, but I took this opportunity to quit Twitter. I didn't delete my account but didn't log back in. One of the best decisions I ever made.
When taken in context it looks like an innocuous tweet but to a twitter censorship algorithm that isn't aware of the context it looks like a command for someone to die.