Is AI better at finding answers than questioning assumptions?

AI is becoming remarkably good at finding patterns, generating ideas, and exploring large solution spaces.

But I wonder whether its greatest limitation may not be intelligence itself.

Many problems are defined by the questions we ask.

A question does more than seek information.

It defines the space in which solutions are allowed to exist.

Throughout history, some of the biggest breakthroughs occurred not because better answers were found, but because someone challenged an assumption that everyone else had taken for granted.

The question changed.

The entire solution space changed with it.

Current AI systems can help us explore possibilities within a given framework, but how good are they at recognizing that the framework itself may be wrong?

Is questioning assumptions fundamentally different from answering questions?

Or is that distinction eventually going to disappear as AI becomes more capable?