Shit storm, if AI is trained on the moral conjectures of arm chair psychologists and philosophers.

Morality is NOT a game of virtues. Morality CANNOT BE DETERMINED by choosing who dies in a runaway railcar scenario.

The fundamental a priori of morality lies in protecting the "innocent" (inculpable) from violation.

Out technology should be prepared to fail gracefully, at every expense of itself. Not to calculate virtue point values and moderate sums. These are contrivances which do not reflect morality in any real world circumstance.