AI-Assisted Collatz Conjecture Disproof Fails, Exposing Critical Lean 4 Kernel Bug
An AI-assisted attempt to disprove the Collatz conjecture using the Lean theorem prover failed, but it successfully exposed a critical soundness vulnerability in Lean's kernel. The Lean development team quickly patched the issue with the release of Lean 4.32.2. Soundness bugs in major proof assistants like Lean are extremely rare and highly critical, as they undermine the fundamental trust in computer-verified mathematics. This event highlights how AI-generated proofs can stress-test formal verification systems and uncover hidden compiler or kernel flaws. The vulnerability resided in the kernel's handling of nested inductive types, where it failed to check "phantom type parameters," allowing the system to accept the logical contradiction "False." The bug was reported by researcher Kiran Gopinathan and patched by the Lean team within approximately one hour.
## BACKGROUND
Lean is an open-source proof assistant and programming language based on the calculus of inductive constructions, widely used for formal verification in mathematics and computer science. The Collatz conjecture is a famous unsolved mathematical problem stating that repeating a simple arithmetic operation on any positive integer will always eventually reach 1.