Create a system that can detect when ai is trying to break the codes or when ai is trying to do something that is bad
