Artificial Intelligence presents improved solutions to a given problem that humanity considers needing of improvement. If enough time passes by and enough intellectual effort is put in, AI will get closer and closer to a near perfect solution to the problem. At that stage, two options will be presented to society:
1) Humanity is forced to adopt that near-perfect AI-powered technological solution to the problem. In that case, AI is making humans chose something they may not want to accept, and thus humanity will spend energy trying to get out of such imposed technological solution keeping it away from perfection.
2) We are free to choose to ignore such near-perfect AI-powered technological solution. For that freedom to be tested and ensured, some humans will decide not to adopt that AI solution. Therefore, it will be a solution that a portion of the humanity will not want to take, and thus unable to reach perfection.
Thus, the only way a perfect AI-powered technological solution can be achieved is by making people believe that some individuals are choosing to be out of it [despite these individuals don’t exist], or by creating an imaginary set of revolutionaries that fights against the technology.