Mathematicians Fear AI Is Redefining Their Profession
Cornell professor Steven Strogatz describes the recent surge in AI-driven mathematical proofs as a terrifying shift, where corporate speed now outpaces human expertise and traditional credit systems.
Steven Strogatz, a professor at Cornell University, admits to feeling genuine fear when discussing the latest capabilities of artificial intelligence in mathematics. While the technical achievements are impressive, he argues that the human cost is significant. The rapid integration of AI into high-level research is creating a rift between traditional academic values and the aggressive timelines of technology companies.
This tension was highlighted last week when OpenAI announced it had solved a 90-year-old math problem worth a one-million-dollar prize. The solution, which remains pending independent verification, was developed using tens of thousands of AI agents. However, the announcement sparked controversy after another mathematician claimed OpenAI rushed to claim credit for work that overlapped with his own research.
Corporate Speed Outpaces Academic Credit
The dispute over the Navier-Stokes problem is seen by experts as a symptom of a broader race for headlines. Strogatz notes that corporate labs are eager to showcase their models' capabilities, often ahead of major financial events like initial public offerings. This creates a pressure to announce results quickly, sometimes before the academic community has fully vetted the work or agreed on who deserves recognition.
Similar issues have arisen with other recent claims. Anthropic reported that its model proved thousands of small theorems while formalizing a proof of Fermat’s Last Theorem. These developments suggest that AI is not just assisting but actively competing in the most abstract corners of pure mathematics, where human intuition has traditionally held the advantage.
Mathematicians Question Their Own Relevance
For working mathematicians, the emotional impact is profound. Alex Townsend, a collaborator of Strogatz, recently used AI to solve a decades-old problem in numerical linear algebra. He acknowledges that without this technology, the volume of work required would have been prohibitive. However, he also expresses a deep sense of loss, feeling that his peak professional skills are no longer unique or necessary for breakthrough work.
Townsend describes the shift as unsettling, noting that he no longer feels like he is at the frontier of knowledge. Instead, he feels like a user of a tool that surpasses his own abilities. This perspective is becoming common among experts who have dedicated their careers to pushing the boundaries of human understanding.
The Trade-Off Between Speed and Understanding
Strogatz likens the current situation to a horror movie, where an incomprehensible force is closing in. He warns that the year 2026 may be remembered as a pivotal moment for mathematics, either a year of wonders or a year of horrors. The core issue is that AI systems are driven by processes that are not fully understood by the humans using them.
According to reporting from GN technics/ai (en-US), the stakes are high. While AI accelerates discovery, it may also erode the deep understanding that comes from years of human struggle with complex problems. The field now faces a difficult choice: embrace the speed of machine-generated proofs or risk falling behind in a discipline that is rapidly changing.
