AI is not intricate If/Then programming.
I think this article is a bit naive on what AI is and how it works. And given another 100 years AI will easily surpass humans.
My real fear is that as seen in many sci-fi novels/movies. An AI that thinks it must protect humans will continue to evolve itself until it determines the best way to protect us is to enslave us.
That is my interpretation, having worked with the mathematics of artificial neural networks quite a bit.
It is also my opinion that true intelligence requires self-awareness/consciousness, and there is a quantum mechanical component to self-awareness/consciousness that can never be achieved by software alone.
Such AI is well behind the authoritarians, benign or otherwise, who control from behind the facade of a giving and essential government. They are significantly aided by those politicians who are satisfied simply by fulfilling personal needs.
Of the two, authoritarians and AI, only the former can guarantee a sufficient power source, and the latter can be allowed to die.