Mathematicians Fear AI’s Brute-Force Approach Undermines the Art of Proof

Artificial intelligence has already disrupted creative fields ranging from music composition to acting, but mathematics is now facing its own upheaval. Following OpenAI’s announcement that it utilized thousands of agents to resolve the decades-old Navier-Stokes existence and smoothness problem, some researchers are concerned that the discipline is losing its human touch.

“The artists and the musicians have already gone through this,” noted Juspreet Singh Sandhu, a mathematician at Colorado State University. The incident highlights a growing tension between computational power and the traditional, deliberate methodology of mathematical discovery.

Contrary to the stereotype of math as a race to the correct answer, the field is deeply rooted in creative exploration. The development of new concepts often mirrors artistic endeavors, such as inventing a new game or puzzle. In contrast, OpenAI’s method relied on brute force, effectively shortcutting the thoughtful process that defines the discipline. Critics argue this approach threatens to undermine genuine human understanding of the underlying principles.

This perspective aligns with the views of G. H. Hardy, the English mathematician who, in his 1940 essay *A Mathematician’s Apology*, described mathematicians as “makers of patterns” much like painters or poets. Written during World War II, Hardy’s essay advocated for pursuing mathematics for its intrinsic beauty rather than practical application, citing Carl Friedrich Gauss’s praise for number theory as the epitome of beautiful, “useless” mathematics. While Hardy was correct about the artistic nature of math, his dismissal of utility proved premature; number theory later became foundational to modern encryption protocols used to secure emails and banking transactions.

However, the recent proof regarding the Navier-Stokes equations—which describe the flow of viscous fluids and are used by engineers to model airflow for aircraft design—falls squarely into the category of “useless” in terms of immediate application. According to Jared Speck, a mathematician at Vanderbilt University, the primary interest for mathematicians lay in the equations themselves rather than their engineering utility. The pursuit was driven by intellectual curiosity, not practical necessity, raising questions about whether an AI-generated solution can capture the essence of that specific form of inquiry.