You have put your finger on the crucial question: whether AI research has killed everyone in the past. How overcareful would a person have to be to worry about its happening in the future if it's never happened in the past?