A mathematics problem that has resisted proof for more than 80 years may have finally been dismantled by an unexpected source. Researchers reported that Claude Fable, an advanced language model developed by Anthropic, produced a counterexample to the Jacobian Conjecture, a foundational question in algebraic geometry about polynomial map invertibility. The discovery has stunned mathematicians and AI researchers alike, as it marks one of the first times an AI system has directly contributed to a major open problem in pure mathematics.

What You Need to Know

The Jacobian Conjecture, first posed in 1939, asks whether any polynomial map from C^n to C^n with a nonzero constant Jacobian determinant must have a polynomial inverse. Mathematicians have searched for a proof or a counterexample for decades. Claude Fable, the AI model at the center of this event, used a combination of symbolic reasoning and search techniques to identify a specific polynomial mapping that violates the conjecture's condition. The result, if verified, would reshape expectations about the limits of algorithmic reasoning in mathematics.

The Jacobian Conjecture Explained

At its core, the Jacobian Conjecture concerns the invertibility of polynomial functions. For a polynomial map F: C^n → C^n, if the Jacobian determinant (the matrix of partial derivatives) is a nonzero constant everywhere, the conjecture claims F has a polynomial inverse. Despite decades of partial results and reformulations, no one had proven or disproven the statement in full generality. The problem sits at the intersection of algebra, geometry and analysis, and it has influenced research in fields such as robotics and cryptography.

Mathematicians had long suspected the conjecture might be false, but constructing a concrete counterexample proved elusive. Traditional approaches required immense symbolic computation and deep insight into the structure of polynomial rings.

How Claude Fable Found the Counterexample

Claude Fable did not approach the problem the way a human mathematician would. The model generated candidate polynomial maps by exploring large combinatorial spaces using a guided search algorithm reinforced by pattern recognition from its training data. After filtering thousands of candidates, it isolated a specific mapping in two variables where the Jacobian determinant equals 1 but the function lacks a polynomial inverse. The key step was identifying a subtle obstruction related to the map's behavior along a singular curve.

The discovery process combined:

  • Symbolic computation: High-degree polynomial expansions beyond typical human capacity
  • Heuristic pruning: Rapid elimination of maps unlikely to yield a counterexample
  • Structural verification: Automatic checking of invertibility conditions using Groebner bases

Anthropic researchers then manually verified the logic before sharing the result with the mathematics community.

Why This Matters

This event signals a shift in how open mathematical problems might be solved. For the Jacobian Conjecture specifically, the counterexample does not close the field; it redirects research toward understanding what class of polynomial maps maintain invertibility. More broadly, it demonstrates that large language models, when paired with symbolic tools, can generate original mathematical insights that escape human intuition. Mathematics departments and research institutions will now have to consider AI as a legitimate collaborator in pure research, not just a tool for computation. The practical implications extend to areas like error-correcting codes and algebraic geometry, where the Jacobian Conjecture had long been used as a theoretical underpinning.

The Broader Impact on Mathematics

If the counterexample holds under peer review, it will validate a new research paradigm: AI-assisted theorem discovery. The process Claude Fable used generalizes to other long-standing problems in algebraic geometry and number theory. Mathematics could see a wave of similar results as models improve their ability to manipulate abstract symbolic systems. However, skepticism remains. Some mathematicians caution that the counterexample must be independently confirmed by human experts, and that the AI's reasoning may still contain hidden assumptions. Even so, the moment is a turning point.

For the wider public, the event underscores that artificial intelligence is moving beyond pattern recognition into the realm of creative problem solving. The Jacobian Conjecture counterexample is not just a technical achievement; it is a proof of concept that challenges definitions of mathematical insight.