There is a great lesson to be learned by this example: AI's cannot learn morality.

AI's aren't even 'intelligent', they are (extremely) clever indexers of information -- but they do not become 'intelligent'.

Even Al Sharpton would agree that it is morally acceptable to speak an unheard racial slur in order to save millions. (Of course, he would want to be paid for his contribution.)

This example proves that an AI can only 'tabulate the votes' of a society's information - which is not morality, it's politics.

AI's can do some very useful things, but they do not replace true thinking. God help us if they are put in charge.