{"resourceId":"papex-eth-refutation-claims-lean-checked-no-expert-verdict-20261031","versions":[{"version":"external-cf68d476b41b8ce2059130955ab26d7be9977064552a13fe195a16d6d7d40eae","resource":{"id":"papex-eth-refutation-claims-lean-checked-no-expert-verdict-20261031","title":"Claude's claimed refutation of the Exponential Time Hypothesis: Lean proofs replay, but no outside expert verdict yet and follow-up AI-written papers already build on it","organization":"Papex (covering Anthropic and OpenAI claims)","sector":"Academic research; theoretical computer science","geography":"Global","publishedAt":"2026-10-31 (sources checked Oct. 30)","publicationDate":"2026-10-31","eventDate":"2026-10-30","sourceName":"Papex","sourceLabel":"Independent research-integrity reporting","sourceUrl":"https://papex.org/blog/eth-refutation-what-has-been-checked","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["knowledge-work","governance-procurement","developers-agents"],"finding":"As of Oct. 30, no one outside Anthropic or OpenAI had published a verdict on either lab's claim that its AI refuted the Exponential Time Hypothesis. Independent replays confirm the Lean proofs check against their own formal statements. Whether those statements mean \"ETH is false\" is a human reading not yet done, and the organizers say their involvement is not certification. Follow-up preprints, several mostly AI-written, already treat the claims as imported theorems.","sledRelevance":"University research offices, computer science departments and research-integrity officers face AI-produced results, and AI-written follow-up papers, before expert verification exists (cross-tagged research because it concerns how institutions treat machine-checked but unreviewed AI research claims). Why now: published Oct. 31, not covered in the Oct. 31 edition and not archived; the open examination is ongoing with a meeting proposed before the end of 2026.","evidence":"Independent synthesis of public examination materials, Lean replays and preprints; no expert verdict.","architectureImplications":"Formal verification confirms a proof matches its formal statement, not that the statement matches the paper; research AI pipelines need a human statement-review step.","governanceImplications":"Research-integrity and authorship policies should address AI-generated results and AI-written follow-up work that cite unverified claims.","securityPrivacyImplications":"No direct data-security effect.","caveats":"Fast-moving; status may change. Papex shows each claim's status as Claimed and gives no verdict of its own.","streamIds":["anthropic","research"],"roles":{"sales":"For research universities, the sales point is rigor. Claude's claimed ETH refutation shows what AI research tools may produce, but as of Oct. 30 no outside expert had confirmed it, and the organizers say their involvement is not certification. Position Claude for research with verification workflows, not as a discovery engine. Ask how research offices handle AI-produced claims and AI-written papers. Do not cite the ETH result as proven.","engineering":"For engineers supporting research computing, the lesson is that formal tools check proofs against formal statements, while a person still has to confirm the statement says what the paper claims. Pipelines using Claude for proofs or code should include human statement review, reproducible builds and recorded provenance. Pin toolchain versions so replays are repeatable, and keep the statement file short enough for a person to read against the paper.","delivery":"Delivery can support research offices with a policy and workflow for AI-produced results: provenance labels, a human review step for formal statements, and guidance on citing unverified AI claims. Acceptance criteria are proposals: policy approved by the research-integrity office and the workflow piloted in one department. Track the examination outcome before using the result in training material, and update guidance if the examiners report a gap or correction."},"roleDetails":{"sales":{"customerProblem":"Research universities face AI-produced results and AI-written follow-up papers before expert verification exists.","stakeholders":["Vice president for research","Research-integrity officer","CS department chair","Library scholarly communications"],"discoveryQuestions":["How does the institution label AI-generated research contributions?","Who reviews formal statements in machine-checked proofs?","Are faculty citing unverified AI results?"],"valueHypothesis":"Interpretation: clear verification and provenance practice can reduce reputational risk as AI research tools spread.","potentialEngagement":"A research-AI provenance and verification policy workshop.","unsupportedClaims":["That the ETH refutation is correct","That Lean checks establish the paper's theorem","That Claude reliably produces valid research"]},"engineering":{"fit":"Relevant to research computing using Claude for formal proofs or code.","architectureIntegration":"Add reproducible formal-verification builds and human statement review to research pipelines.","prerequisites":["Lean and Mathlib toolchain","Compute for replays","Named statement reviewers"],"constraints":["Machines cannot check that a formal statement matches the intended theorem","Expert review capacity is scarce"],"security":"No direct security concern; record provenance of AI-generated artifacts.","proofOfValue":["Replay a published Lean development in campus research computing","Document a statement-review checklist"]},"delivery":{"work":"Draft and pilot a policy and workflow for AI-produced research results.","dependencies":["Research-integrity office","Department leads","Research computing"],"ownership":"Vice president for research office.","skillsAdoption":"Training on what machine-checked proofs establish.","governanceCheckpoints":["Policy approval by research-integrity office"],"acceptanceCriteria":["Proposal: policy approved","Proposal: workflow piloted in one department","Proposal: guidance published on citing unverified AI results"],"risks":["Claims later fail expert review","Policy lags tool adoption"]}},"retrievedAt":"2026-11-02T14:36:45Z","enrichedAt":"2026-11-02T14:44:10Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Faculty and graduate students need training on what machine-checked proofs do and do not establish.","procurementImplications":"Research-AI purchases marketed on discovery claims should be assessed on verification practice, not headline results.","operatingModelImplications":"Institutions may need review routes for AI-produced results before they are cited, taught or deposited.","sourceVerification":{"openedUrl":"https://papex.org/blog/eth-refutation-what-has-been-checked","referenceExcerpt":"no one outside the two companies has published a verdict on either claim.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}