# AI Outperforms Human Mathematicians

AI is surpassing humans in formalizing mathematical proofs.

- **Source:** [Hacker News: Front Page](https://brifing.si/source/b07aa772-b1a0-4a76-aa18-719c6f9e2dde)
- **Published:** 2026-07-20
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://xenaproject.wordpress.com/2026/07/20/human-mathematicians-are-being-outcounterexampled/>
- **On Brifing:** https://brifing.si/s/4faee451-4729-4458-8824-44077352e246

## Introduction

Artificial intelligence is surpassing humans in formalizing mathematical proofs. ChatGPT's disproof of Erdős' Unit Distance conjecture marked a turning point. Verified by human mathematicians, this proof highlights AI's role in mathematical discovery.

## ChatGPT's Achievement

In May 2026, ChatGPT disproved Erdős' Unit Distance conjecture. Using a 1960s number theory theorem, the proof stunned the math world. Early access verifications by human mathematicians showcased AI's capabilities.

## Formalization with Lean

The proof's formalization in Lean demonstrated AI's potential to automate mathematical processes. Logical Intelligence's systems autoformalized ChatGPT's paper into Lean, proving AI can handle complex mathematical structures.

## Global Class Field Theory

AI began formalizing complex topics like global class field theory. Boris Alexeev used OpenAI's Sol model to fully formalize Erdős' counterexample, showcasing AI's ability to process 100+ page theorems.

## AI Code Generation

The Sol model generated 1.2 million lines of Lean code in three weeks. Compared to mathlib's 2.3 million lines over nine years, AI's speed is remarkable, signaling accelerated mathematical advancements.

## Security Concerns

Trust in AI-generated code is debated. Lean's programming nature risks malicious code execution, but sandbox tests confirmed AI's ability to validate nontrivial mathematical theorems.

## Fermat Workshop

The Formalizing Fermat workshop highlighted AI's role in mathematical formalization. Tools from Logos Research and others offered AI-assisted formalization, with participants experimenting using Claude Max and ChatGPT Pro.

## AI Counterexamples

AI can detect errors in mathematical texts. Logos' tool identified a false claim in an LLM-generated PDF and provided a counterexample, demonstrating AI's potential in mathematical validation.

## Mathematical Discoveries

AI is accelerating mathematical discoveries. Models like ChatGPT formalize complex theorems, aiding human mathematicians. This development could reshape the future of mathematical research.

## AI-Human Collaboration

Collaboration between AI and human mathematicians boosts efficiency. AI's rapid formalization allows humans to focus on creativity, creating a symbiotic relationship that accelerates progress.

## Future Predictions

AI-assisted mathematical formalization will become standard. Large-scale AI-generated mathematical developments seem inevitable, potentially revolutionizing the nature of mathematical research.

## Conclusion

AI is surpassing humans in formalizing and discovering mathematical proofs. Tools like ChatGPT are making strides in solving and formalizing complex theorems, possibly heralding a new era in mathematics.

## Sources covering this story

- https://xenaproject.wordpress.com/2026/07/20/human-mathematicians-are-being-outcounterexampled/ — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._