Substitute the word 'educator' for 'leader' and the entire argument is exactly the same. If enjoined to discover what a population actually wants and to then implement it, it seems to me that robots could soon do a better job than our leaders. Not yet, though. The hard part is the first part - determining what people actually want. That's what could be solved with computation, at least in theory. Human leaders fail in the second part: implementing it. Too often, human leaders serve some other agenda, one that acts against the interests of the population.