It's perfectly reasonable to say that AIs are potentially dangerous. That seems to me like a statement of fact.