What was already known
Earlier unconditional attacks covered two rounds or special cases; an arbitrary-round polynomial-query attack relied on a conjecture.
AI involvement
The authors credit GPT-5.6 Sol Ultra with discovering the proof in one conversation. Longcheng Li, Qian Li, Xingjian Li and Qipeng Liu verified and wrote the final paper.
The authors explicitly attribute discovery of the complete proof to a one-shot model conversation.
Importance
Removes that conjecture for perfectly complete protocols in the stated model.
Limitations
The four human authors report independently checking every statement and proof, then simplifying and refining the argument. The result does not cover imperfect completeness, entangled setup or quantum communication.
Technical verification
All four authors explicitly report independently verifying every statement and proof of the final theorem; this is a source-reported complete human check, not an AIProofWatch reproduction.
Formal checking
Theorem 3.1, the perfectly complete finite-round QCCC key-agreement eavesdropper with O((q_A+q_B)^5) classical queries; Independent author verification of every statement and proof, followed by simplification and refinement
- Checker
- Longcheng Li, Qian Li, Xingjian Li and Qipeng Liu
- Scope
- Theorem 3.1, the perfectly complete finite-round QCCC key-agreement eavesdropper with O((q_A+q_B)^5) classical queries
- Basis
- Reported by the linked checker
Checking a formal statement does not establish that it matches the original problem.
Human understanding
The authors simplified and rewrote the argument and explain its transcript factorization, disjoint polynomial supports, separator decision tree and binary search.
Correction and publication history
Removes that conjecture for perfectly complete protocols in the stated model.
Read this published revision