Yes, they are. So what would the correct policy be if we have no reliable model? Well as all existential threats should be treated, with extreme risk aversion.