So who is ultimately responsible? AI isn’t something like an unfortunate mechanical failure (though even than someone bears responsibility). If the government is using AI for targeting military attacks, it sure better understand every single risk associated and be ready to swallow every ounce of responsibility. You don’t get to say “AI messed up, oops” and move on. You told the machine what to do, so you’re responsible. Every one of those school girls’ blood is on your hands.