People believe that one day AI will totally eclipse human intelligence, but there is already a mathematical proof called the incompleteness theorem which indicates that there may exist certain fundamental differences that no AI could overcome.
In short the incompleteness theorem states that no system can prove its own consistency by using the logical deductions that exist within the axioms of that system. This can be interpreted to mean that AI are programmed within a box that has a set amount of rules can cannot make leeps of logic outside that boss.
Humans however, are somehow able to transcend this limitation by thinking outside the box. Humans beings can make logic jumps to entire new planes that are unconnected to the previous one. This implies a fundamental difference between biological thinking and machine thinking.
This fundamental difference between human thinking and machine thinking can be inferred as intuition, and it is something robots can never have because it is not quantifiable.
This is not only speculatory observation, but there is some actual evidence to back up these claims from a very prominent figure, Albert Einstein. Einsteins developments of special and general relativity required a whole new way of thinking. You could take an infinitely powerful AI, plug the physics of the times into that AI, have it deduce all possible conclusions from those parameters, and it would still not arrive at Einsteins equations because they required a completely new way of thinking.
Here is a short documentary arguing this point.