Dear Aaron,

 for including ethics into AI courses, you may consider taking ethical principles and illustrating those in terms of current AI applications students can relate to.  Some principles are listed below and I am sure can find applied cases. I think the subject can be treated in non-ideological terms.

Ethics in the business arena can be defined in terms of knowing what is good and what is not good. The choices between the two are many, are in conflict, and involve personal and corporate collective consciences in a variety of cultural contexts. 

Conscience applies to what happens before, during and after an Artificial Intelligence algorithm has been deployed. It is often a battle between economic interests and values. 

Rights refer to the natural person; the right to work, to property, to happiness and fellowship. Defining documents are the U.S. Constitution’s Bill of Right, the United Nations’ Declaration of Human Rights, and the Geneva Convention for the conduct of war. Artificial Intelligence must not violate those rights.

A Calling aims to contribute to an improved way of life for natural persons. In Artificial Intelligence this may involve shifting menial- and dangerous tasks from humans to robots. 

A Measure of Wisdom. Although business managers in corporations are not presumed to be per se unethical, one should start with an understanding that “power corrupts, and absolute power corrupts absolutely”. (Lord Acton, 1887).  Given an opportunity for an unethical decision, it is presumed that somebody, but not all, will “rise “to that occasion.

The Obligation to avoid common unethical outcomes in commerce rests with management. Examples for unethical applications are contempt for classes of citizens, to take advantage of the disadvantaged, constraining zeal for winner-take-all. 

Limits and Boundaries: If the ethical good in AI- commerce is to give the people what they need and want; then the question arises of whether “too much good is bad”. For instance, if robots can free humans from having to work to the extent that the average person needs to work only three days out of five without loss of income, then how would the average person use the resulting free time during a short transition period?

Unintended Outcomes are a large risk in Artificial Intelligence and often the result of negligent oversight or absence of it. Oversight negligence implies a lack of human control once an algorithm is deployed.  For example, a physician may rely on a prescription algorithm that is 99% accurate but may harm 1 in 100 patients. The responsibility must remain with the physician, not the algorithm. 

Good Economic rules are efficiency (not being wasteful), effectiveness (correctly addressing a given problem), and productivity (benefits outweighing costs).  

Frank