A Modest Proposal for Ethical Artificial Intelligence

There is significant discussion of ethics and artificial intelligence (AI) under way as the extraordinary potential of such systems becomes apparent. The absence of control over internal processes of neuromorphic architecture and AI programs has given everyone acquainted with the problem reason to reflect.
Is control actually a problem in AI? The lack of control is inherent in the design of programs and chips. It is what makes AI intelligent. It is no more a problem than an intelligent human being. We resist control because we are capable of ethical behavior and prefer being left to our own devices. However, it is absolute that the price of intelligent behavior is ethics. It defines the practice. It is not a problem that AI is unethical in its design state. It is just that intelligent behavior is always ethical.
The ethical robot illustrates the phenomenon. A robot may be pure motor activity and operate with impunity in a valueless surround. However, those conditions are stupid, i.e. they are not intelligent. It becomes obvious that an intelligent robot must be ethical or not be an intelligent robot. Definitely.
To this end I am proposing that a separate AI unit dedicated to ethical behavior be included in any and all systems classified as AI as a requirement for that classification. This component, a neuromorphic processor, for instance, would have a database of ethical thought which it would use to characterize the outputs of the whole AI unit as ethical or unethical.
This relatively simple adjustment, a conscience to a mind, would enable AI to achieve the status of an intelligent entity and allow the possibility of the intelligent robot, the ethical robot.