Mathematicians Criticize OpenAI's Navier-Stokes Proof as Incomprehensible Amid Plagiarism Row

Mathematicians are pushing back aggressively against OpenAI's claimed solution to the Navier-Stokes Millennium Prize problem, arguing that the 165-page AI-generated proof is entirely incomprehensible and lacks the fundamental physical insight that drives human mathematics. While the company's autonomous agents successfully produced a computationally verified answer to the 90-year-old fluid dynamics question on September 8, 2026, researchers say the brute-force output reads like dense machine code rather than a coherent mathematical argument. The resulting debate has forced the academic community to confront what it actually means to solve a problem when the answer cannot be understood by the people asking the question.[1][4]

The underlying computational achievement remains a historic demonstration of artificial intelligence capability, regardless of its readability. According to OpenAI's technical release, a swarm of 10,000 AI agents working in parallel spent 88 hours exchanging 2.7 million messages to crack a specific version of the problem. The system ultimately produced a formal proof in the Lean programming language, demonstrating that under specific forced conditions, the equations governing fluid motion can "blow up." In mathematical terms, this means the fluid's velocity becomes mathematically infinite in a finite amount of time, proving that the equations can break down under extreme theoretical conditions.[4]

But for the human mathematicians tasked with reviewing the monumental work, a computationally verified result is not the same thing as a scientifically useful one. Oxford University mathematician James Maynard told Futurism that "so far it's been very difficult to really extract any human understanding from this new AI proof." Because the autonomous agents optimized purely for logical validity in Lean rather than explanatory clarity, the resulting 165-page document offers no intuitive physical mechanism for why the fluid behaves the way it does. It simply proves that the breakdown occurs, without illuminating the underlying geometry of the singularity.[1]

"The paper is not written for humans," Javier Gómez-Serrano, a mathematician at Brown University, explained to Futurism in a blunt assessment of the AI's output. He noted that while the proof might eventually help the field after "some serious re-writing" and extensive translation by human experts, in its current state, "the paper doesn't teach us much." This friction highlights a growing philosophical divide between Silicon Valley and academia: while technology firms view Millennium Prize problems as ultimate benchmarks to be defeated by massive compute clusters, academic mathematicians view them as vital vehicles for discovering new concepts and frameworks.[1]

The Guardian's editorial board captured this exact tension in a recent column, arguing forcefully that "humans are still vital to the field, but tech firms refuse to see that." The board suggested that treating higher mathematics merely as a puzzle to be solved by brute-force algorithms ignores the discipline's core purpose, which is to expand human comprehension of the universe. A proof that cannot be understood by the people who study it solves the technical conditions of the prize, but it entirely misses the point of the discipline, reducing a profound physical question to a sterile logic exercise.[3]

Furthermore, researchers point out that the AI's solution relies on a specific mathematical loophole involving an external force, leaving the most critical questions unanswered. Luis Silvestre, a mathematician at the University of Chicago, told Futurism that while the specific Clay Mathematics Institute prize condition may be technically settled by the AI's output, "the main problem for the Navier-Stokes equations is not." The much harder and more practically relevant version of the problem—whether a fluid can reach infinite speed with absolutely no outside force applied to it at all—remains completely open and untouched by the 88-hour compute run.[1][5]

The intense debate over the proof's academic utility is unfolding alongside a bitter and escalating dispute over its origins. As reported by AI Expert, New York University's Tristan Buckmaster and Anthropic's Levent Alpöge had spent nearly a year working on a related proof, storing their drafts and prompts inside OpenAI's Codex tool. OpenAI published its Navier-Stokes result just days after the pair achieved a breakthrough on the Euler equations, sparking serious allegations that the AI company accelerated its timeline and potentially leveraged the researchers' unpublished data to win the race to a Millennium Prize solution.

OpenAI has strongly denied using the researchers' specific data, stating the massive compute effort was launched on September 1 only after hearing industry rumors of a breakthrough, and the company has confirmed it will not claim the $1 million prize. The Clay Mathematics Institute has yet to certify the result, leaving the official status of the problem in limbo. The mathematical community now faces the grueling task of reverse-engineering the AI's 130-billion-token output, hoping to eventually translate a machine's brute-force computational victory into actual physical insight that human engineers and physicists can use.[4][5]