Has Mathematics Become a Thing of the Past?

After OpenAI announced a proposed solution to one of the million-dollar Millennium Prize Problems in mathematics, a second announcement came that was even more startling to the scientific community: more than 700 AI-generated mathematical manuscripts were published and made available for examination. Between admiration for these achievements and concern about their implications, one question deserves particular attention: are we putting human understanding at risk?

Yet these remarkable achievements do not automatically translate into advances in human understanding. Traditionally, when mathematicians arrived at new solutions or proofs, they would present their findings at seminars, explain their reasoning and share their newfound understanding. Others could then build on those achievements, sometimes applying the same insights in neighbouring fields.

This raises the question: has mathematics become a thing of the past? Indeed, what is the point of learning and understanding it if machines can now solve problems that have defeated scientists? And what is the role of the mathematician if artificial intelligence can produce solutions and proofs faster than humans can?

These questions may seem provocative, but they deserve reflection. These developments have placed mathematics in the spotlight, while many other disciplines and professions are facing similar concerns. Everyone is trying to find their place in a world where the capabilities of machines continue to advance, and where fear is mixed with optimism and the desire to adapt effectively.

Let us imagine the world ten years from now. Artificial intelligence has advanced to the point of formulating new theories in physics and mathematics, but human understanding has not kept pace. We ask it to design a nuclear reactor or an aircraft based on those theories, and it presents us with a design unlike anything we have seen before.

The problem is that we do not understand why it works, or when it might fail.

Would we board that aircraft? Would we agree to build that reactor near our cities simply because the system that designed it assures us that it is safe?

When human lives are at stake, the value of understanding becomes clear. We need to know why something works, where its safety limits lie, and how to detect and address potential failures. Even if another AI system reviewed and approved the design, the question would remain: do we ourselves have the ability to judge?

This is where scientists and engineers come in. Their work goes beyond producing answers to interpreting them, testing them and connecting them to reality. Artificial intelligence may generate results faster than we can absorb them, which only increases our need for people who can turn those results into knowledge that we can understand and use responsibly.

This brings us back to our first question: why do we learn mathematics? Because its value extends beyond obtaining answers. It teaches us how to think, formulate questions, understand relationships and distinguish what can be trusted. Obtaining an answer, however impressive, does not give us the ability to make sound decisions about it. It is our understanding of the problem that allows us to judge how reliable a solution is, where its limitations lie and what it might overlook.

The gap may widen if we become accustomed to receiving answers without making the effort to understand them. The danger is that our human capacity to review what machines produce will decline, and that we will gradually lose the knowledge needed to question and correct them. This is not necessarily a distant prospect.

These developments open up enormous possibilities, but they also place a great responsibility on us. AI companies are racing to develop increasingly capable systems. Alongside them, the scientific community must develop the tools needed to examine and understand their outputs, while our education systems must equip people with the knowledge and skills to use these technologies critically and effectively.

The role of humans in this landscape may change, but our responsibility for the decisions we make remains. Perhaps the greatest challenge in the coming years will be preserving our ability to understand and choose while increasingly relying on an intelligence that surpasses us in certain tasks.

Mathematics has not become a thing of the past, but our relationship with it, and with knowledge in general, is entering a new phase — one that requires us to ask the right questions, at the right time.