FB does scan conversations for keywords, but it doesn't seem to understand the context of the conversation. My one nephew is a FB star and had used the phrase "I'd have to kill you" in one of his posts. He was just joking with a friend, but he got suspended for a month because of it.
This is where it's retarded, or malicious. "Everyone" knows that "I'd have to kill you" is a cliche. But what's genuinely malicious is that you don't have to suspend someone for a month automatically just because they say something like that. In a decent world, the platform would ask the user for a "please explain" before taking any actual action. The fact that they don't is proof that they are not decent.