For centuries, scientific progress has depended not only on human curiosity, but also on the tools we have created to explore the world. The telescope allowed us to see further into space. The microscope revealed worlds too small for the human eye. Computers transformed what scientists could calculate. Artificial intelligence may now be opening another door: helping us explore questions that have remained beyond our reach.
A recent development in mathematics offers an intriguing example. As Science News reports, an AI system has produced a possible solution to the Navier–Stokes problem, one of the seven Millennium Prize Problems established in 2000. These are among the most difficult unanswered questions in mathematics, each originally carrying a $1 million prize for a verified solution.
Navier–Stokes is concerned with something deceptively familiar: how fluids move. The equations are used to describe phenomena such as airflow around an aircraft, weather and the movement of blood. Yet one fundamental question about them has remained unanswered for around 90 years: can a perfectly regular flow eventually develop a mathematical singularity, in which the velocity becomes infinite?
The proposed AI-generated proof suggests that it can. The argument has also been formally checked using Lean, a computer system that can verify mathematical steps. But mathematicians still need to examine the proof and determine whether it genuinely resolves the problem.
Whatever the final verdict, there is something remarkable about the moment itself.
The Millennium Problems were created to identify questions that could define the mathematical achievements of a generation. For decades, their difficulty represented the limits of what mathematicians could prove. Now, artificial intelligence is beginning to act as a new kind of research tool: not simply calculating faster, but exploring possible arguments, connecting ideas and producing mathematical work that humans can then examine.
And this is not happening in isolation. Around the same time, mathematicians Tristan Buckmaster and Levent Alpöge reported important progress on a related fluid-dynamics problem using AI tools. Their work is another example of a future in which discoveries may emerge from collaboration between human researchers and artificial systems.
That does not make human scientists less important. Quite the opposite. The role may be changing from doing every calculation ourselves to asking better questions, designing new approaches and deciding what the results mean.
Perhaps that is what makes this story bigger than one difficult mathematical problem. Science has always advanced by extending human abilities with new tools. If AI becomes another such tool, the most important discoveries may not be those it makes instead of us, but those that become possible because we can now investigate questions that were previously out of reach.
PhD in Sociology from the University of Barcelona. Early Childhood Education Teacher. Substitute Teacher at the Universitat de València.