AI and the World's Hard Problems · Quantum physics & quantum computing · arXiv · 2026-10-23
Lean-QuantumAlg: A Framework for AI Agents to Produce Quantum Algorithms that Humans Can Trust
AI agents can now build quantum algorithms and proofs, but machine-checked proofs alone do not guarantee that the intended problem is solved or that resource claims match the verified computation. The authors introduce Lean-QuantumAlg, a Lean 4 framework whose contract binds correctness, success probability, borrowed-state restoration and resource bounds to a single execution. Demonstrated on nine AI-formalized algorithms, from Grover search to quantum attacks on RSA-2048 and P-256, it requires human statement-level review of under 1% of each algorithm's imported source code once the shared framework is audited.