What should happen if AI is socially problematic?

Suppose an advanced AI suggest that the majority of people can improve their life standard and live happier with some changes, but those changes would cause grieve to a tiny amount of very powerful people. That is suppose that with very general assumptions and valid arguments the AI is able to simulate a broad range of futures in all of which the change is a considerable improvement worldwide except for that 0.001 % of the population that have a huge amount of money and power. Also suppose that the most advance AI in our planet is financed by that very rich group, could be possible that the rich group would make some change to avoid reaching the good conclusion?