The rise of artificial intelligence has created excitement, but it has also stirred unease among mathematicians. That tension became especially visible after OpenAI announced an AI-generated solution to the Navier–Stokes existence and smoothness problem, one of the Millennium Prize Problems. The Clay Mathematics Institute later described the problem as “apparently” settled while noting that its evaluation process is deliberately unhurried.
The announcement invites a deeper question. In mathematics, is reaching a solution the same thing as understanding it?
The path creates more than the destination
Leonhard Euler’s exploration of the Seven Bridges of Königsberg is a useful historical example. The question seemed simple: could someone walk through the city and cross each bridge exactly once?
Euler’s conclusion mattered, but the method mattered even more. By stripping the problem down to connections between points, he created ideas that helped establish graph theory. Those ideas now support work in network design, routing, logistics and computer science. The original puzzle produced an entire mathematical language.
An answer can close a problem. A method can open a field.
Knowing that is not the same as knowing why
The Navier–Stokes equations describe the motion of fluids. Settling their famous mathematical question would be a remarkable result, but a formal solution alone does not instantly redesign an aeroplane or perfect a weather forecast. Progress comes when people examine the proof, understand its ideas, test its limits and connect those ideas to other problems.
This is why the journey matters. Rigorous exploration develops intuition. Failed approaches reveal the shape of a problem. New concepts sometimes appear because a person spends years asking not only whether something is true, but why.
What do we preserve?
The phrase “if you do the math” is usually an invitation to follow the numbers and logic for yourself. As machines take on increasingly complex calculations and proofs, we should make sure future generations do not lose contact with that process.
Entrusting every mathematical problem to AI could weaken the curiosity and creative struggle that drive discovery. But AI does not have to end that journey. It can also become a collaborator: suggesting paths, checking reasoning and giving people more room to ask better questions.
The challenge is to build a culture that values both the result and the exploration that produced it. “If you do the math” should still encourage us to participate, because there is so much more to discover between trying to solve a problem and finally reaching the answer.
Context: Read OpenAI’s announcement and the Clay Mathematics Institute’s response.