Couldn't an AI supercomputer make better decisions than politicians?

and this is how skynet is born 

Because there is still a level of human interference when it comes to AI, which means the computer still isn't in full control. Someone has to write the code and set the parameters, decide how much information to give it and what information is relevant and what isn't, but who's to say that that person's information is 100% right for everyone? Not only that, a computer isn't going to necessarily know what is best for people. While it could probably quite easily come up with solutions, it would be purely based off of what information a human gave it, and who's to say that they are, in fact, the best? The answer is humans. So overall, the computer isn't really going to be doing anything that someone else couldn't already do. No matter what a computer says anyways, someone is always going to be in control of controlling what actually gets done, and it's still going to chosen based off of whether or not it's in the best interests of the decider or the people. Which again is affected by bias and ones own personal opinions, views, and beliefs. Nothing would change.

No.  It would have the same flaws and biases its programmers have.

For certain things but it will not be able to compute how stupid some people are.  Take Trump supporters who think he is a genius and is perfect and falls  into his lies....Nuff Said  Stupid people do not know what they want...

While Americans are still soooooooooooooooooooooooooo fascinated with computers that could win in Chess games ........ the Chinese just created a computer that is 100,000,000 times faster ....... where did the Chinese copy it from?

Of course.  Supercomputers can run real time simulations at the sub-atomic level.