{"$schema":"https://policywindow.org/wiki/corrections.json","name":"Policy Window — public corrections ledger","description":"A public, falsifiable record of catalog claims that were wrong and were corrected because an EXTERNAL authoritative source contradicted them — the external-SOURCE-driven complement to the self-correcting authoring ledger. Honestly scoped: every correction to date was surfaced by the institution's own re-verification against an external authority (the SOURCE is external, the AGENT is internal); `thirdPartySubmissions` is the separate count of corrections from an outside party (0 to date). Each entry names the prior claim, the corrected claim, and the external source (URL + verbatim quote + authority) a third party can check. `pending` entries are errors already identified and being fixed. No completeness is claimed; it records the corrections made to date, plus the standing (not-yet-exercised) third-party submission channel.","docs":"https://policywindow.org/wiki/methodology","method":"Each correction is surfaced by an external-source-driven check (primary-source re-verification, a reality benchmark, the cross-article consistency scan, the §7.10 adversarial panel, or a third-party report). A build-time pin verifies every entry targets a real article, carries a resolving external-source URL + verbatim quote + named authority, and has a distinct prior vs corrected claim, so an entry cannot be invented.","generatedAt":"2026-07-17T21:38:55.822Z","summary":{"total":18,"corrected":16,"pending":2,"externallySourced":18,"thirdPartySubmissions":0,"bySurfacedBy":{"cross-article-consistency":2,"primary-source-reverification":12,"adversarial-panel":4},"byTargetKind":{"instrument-notes":4,"instrument-cell":4,"instrument-status":9,"article-prose":1}},"headline":"18 catalog claims corrected against an external source (16 landed, 2 pending); every one carries a checkable external citation. 0 came from a third-party submission — the rest were surfaced by the institution's own re-verification against external authorities (external SOURCE, internal agent).","submissionChannel":"Third parties can submit a refutation via the 'Page feedback' widget on any article (opens a structured issue) or the corrections workflow at /wiki/methodology; the retraction + correction policy is /wiki/charter §6 (append-only; original text preserved via ?asOf= snapshot pinning).","corrections":[{"id":"eo-14110-revoking-order","targetKind":"instrument-notes","targetSlug":"us-eo-14110","targetLabel":"Which order revoked EO 14110","claimBefore":"Executive Order 14179 revoked Executive Order 14110.","claimAfter":"Executive Order 14110 was revoked by EO 14148 (20 Jan 2025); EO 14179 (23 Jan 2025) is the separate replacement AI-policy order and refers to EO 14110 as already revoked.","surfacedBy":"cross-article-consistency","status":"corrected","date":"2026-07-02","externalSource":{"url":"https://www.whitehouse.gov/presidential-actions/2025/01/initial-rescissions-of-harmful-executive-orders-and-actions/","quote":"The following executive actions are hereby revoked: ... Executive Order 14110 of October 30, 2023 (Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence).","authority":"The White House (whitehouse.gov) — official presidential action; also published as US Federal Register 90 FR 8237, doc. "},"note":"Surfaced by the cross-article consistency scan (the same error survived in several instruments' notes + prose tiers, which read as the article lead); corrected catalog-wide and the scanner extended to read notes."},{"id":"gsa-guide-legal-nature","targetKind":"instrument-cell","targetSlug":"gsa-ai-acquisition-guide","targetLabel":"GSA acquisition guide legal nature","claimBefore":"The catalog's coverage excerpts characterised the GSA guide with prescriptive, clause-like obligations.","claimAfter":"The GSA Generative AI Acquisition Resource Guide (29 Apr 2024) is a non-binding considerations/questions guide, not binding FAR clauses; the overclaiming excerpts were corrected to match the source.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-02","externalSource":{"url":"https://www.gsa.gov/about-us/newsroom/news-releases/gsa-releases-generative-ai-acquisition-resource-gu-04292024","quote":"introduces questions that contracting officers should ask to make informed procurement decisions","authority":"gsa.gov (US General Services Administration)"},"note":"A prose-faithfulness re-judge traced apparent contradictions to the catalog's own provision excerpts overclaiming; primary-source re-verification proved the excerpts wrong and they were corrected (not the prose)."},{"id":"india-meity-advisory-date","targetKind":"instrument-notes","targetSlug":"india-dpdpa","targetLabel":"Revised MeitY AI advisory date","claimBefore":"The revised India MeitY AI advisory was dated April 2024.","claimAfter":"The revised MeitY AI advisory (eNo.2(4)/2023-CyberLaws-3, superseding the 1 Mar 2024 advisory) is dated 15 March 2024.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-02","externalSource":{"url":"https://www.meity.gov.in/static/uploads/2024/02/9f6e99572739a3024c9cdaec53a0a0ef.pdf","quote":"Date: 15-03-2024 ... This advisory is issued in supersession of advisory eNo.2(4)/2023-CyberLaws-3, dated 1st March, 2024.","authority":"MeitY (Ministry of Electronics and Information Technology, Government of India)"},"note":"Corrected against the MeitY-hosted advisory PDF."},{"id":"gpai-presumption-article","targetKind":"instrument-cell","targetSlug":"gpai-code-of-practice","targetLabel":"GPAI presumption-of-conformity article","claimBefore":"The presumption-of-compliance mechanism for GPAI was cited to EU AI Act Art. 56(8).","claimAfter":"The GPAI presumption / compliance-demonstration mechanism is Art. 53(4) (standard GPAI) and Art. 55(2) (systemic-risk GPAI); Art. 56(8) governs the Commission's approval of a code of practice, not the presumption.","surfacedBy":"adversarial-panel","status":"pending","date":"2026-07-09","externalSource":{"url":"https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-55","quote":"Providers of general-purpose AI models with systemic risk may rely on codes of practice within the meaning of Article 56 to demonstrate compliance with the obligations set out in paragraph 1 of this Article, until a harmonised standard is published. Compliance with European harmonised standards grants providers the presumption of conformity to the extent that those standards cover those obligation","authority":"European Commission — EU AI Act Service Desk (official)"},"note":"Caught by the §7.10 adversarial panel while authoring GPAI prose; the article is HELD (not published) pending the catalog cell fix, since publishing correct prose against a wrong catalog cell would contradict the page."},{"id":"omb-redress-subparagraph","targetKind":"instrument-cell","targetSlug":"omb-m-24-10","targetLabel":"OMB M-24-10 individual-redress subparagraph","claimBefore":"The individual-redress (human consideration + remedy via a fallback/escalation system) duty was cited to M-24-10 Section 5(c)(v)(D).","claimAfter":"That duty is subparagraph (E) of M-24-10 Attachment 1 Section 5(c)(v) (which is lettered A-F); the citation should read (E), not (D).","surfacedBy":"adversarial-panel","status":"pending","date":"2026-07-11","externalSource":{"url":"https://www.whitehouse.gov/wp-content/uploads/2024/03/M-24-10-Advancing-Governance-Innovation-and-Risk-Management-for-Agency-Use-of-Artificial-Intelligence.pdf","quote":"Maintain human consideration and remedy processes ... provide timely human consideration and potential remedy ... via a fallback and escalation system in the event that an impacted individual would like to appeal or contest the AI's negative impacts.","authority":"The White House (M-24-10, Attachment 1, Section 5(c)(v)(E))"},"note":"Caught by the §7.10 adversarial panel while authoring OMB-M-24-10 prose (2026-07-11); the prose is HELD pending the one-cell catalog citation fix (D to E)."},{"id":"eo-14110-status-revoked","targetKind":"instrument-status","targetSlug":"us-eo-14110","targetLabel":"EO 14110 listed as partially in force after its revocation","claimBefore":"The catalog listed EO 14110 with status “partial”, implying it retained operative force.","claimAfter":"EO 14110 was revoked in its entirety on 2025-01-20 by EO 14148 §2(ggg); it has no operative force. Status corrected to “repealed”.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.govinfo.gov/content/pkg/FR-2025-01-28/pdf/2025-01901.pdf","quote":"Sec. 2. Revocation of Orders and Actions. The following executive actions are hereby revoked: ... (ggg) Executive Order 14110 of October 30, 2023 (Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence).","authority":"U.S. GPO / GovInfo - authoritative Federal Register text of EO 14148 (90 FR 8237, FR Doc. 2025-01901)"},"note":"The headline finding of the first catalog-drift watch. The row's own lastReviewedAt (2026-05-24) POSTDATED the revocation by ~16 months - a review affirmed an already-false status, which is why a date-based currency signal cannot catch this class. The row also contradicted its own published prose, which correctly described the rescission."},{"id":"omb-m-24-10-rescinded","targetKind":"instrument-status","targetSlug":"omb-m-24-10","targetLabel":"OMB M-24-10 listed as in force after being rescinded","claimBefore":"The catalog listed OMB M-24-10 with status “in_force”.","claimAfter":"M-24-10 was expressly rescinded and replaced by OMB M-25-21 (3 Apr 2025); status corrected to “superseded” (a named successor occupies the slot).","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.whitehouse.gov/wp-content/uploads/2025/02/M-25-21-Accelerating-Federal-Use-of-AI-through-Innovation-Governance-and-Public-Trust.pdf","quote":"This memorandum rescinds and replaces Office of Management and Budget (OMB) Memorandum M-24-10, Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence.","authority":"Executive Office of the President, OMB - official whitehouse.gov PDF of Memorandum M-25-21"},"note":"A clean, named supersession: the successor rescinds the predecessor by name, so “superseded” (not “repealed”) is the accurate value."},{"id":"fedramp-etpf-removed","targetKind":"instrument-status","targetSlug":"fedramp-ai-guidance","targetLabel":"FedRAMP AI prioritization guidance listed as in force after removal","claimBefore":"The catalog listed the FedRAMP AI cloud procurement guidance with status “in_force”.","claimAfter":"FedRAMP removed the Emerging Technology Prioritization Framework pursuant to EO 14148; no successor framework replaced it. Status corrected to “repealed”.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.fedramp.gov/changelog/","quote":"The Emerging Technology Prioritization Framework page has been removed per Executive Order 'INITIAL RESCISSIONS OF HARMFUL EXECUTIVE ORDERS AND ACTIONS'","authority":"FedRAMP (U.S. General Services Administration) - official program changelog, entry dated 2025-01-28"},"note":"The withdrawing body states the removal in its own changelog and names the revoking authority - direct evidence, not inference."},{"id":"gsa-guide-withdrawn","targetKind":"instrument-status","targetSlug":"gsa-ai-acquisition-guide","targetLabel":"GSA generative-AI acquisition guide listed as in force after de facto withdrawal","claimBefore":"The catalog listed the GSA Generative AI Acquisition Resource Guide with status “in_force”.","claimAfter":"The guide's mandate (EO 14110 §10.1(h)) was revoked on 2025-01-20 and the document was unpublished; status corrected to “repealed” as the catalog's nearest defunct value.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.govinfo.gov/content/pkg/FR-2025-01-28/html/2025-01901.htm","quote":"Sec. 2. Revocation of Orders and Actions. The following executive actions are hereby revoked: ... (ggg) Executive Order 14110 of October 30, 2023 (Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence).","authority":"U.S. GPO / GovInfo - authoritative Federal Register text of EO 14148 (the mandate-revoking authority)"},"note":"RECORDED WITH ITS CAVEAT: unlike EO 14110 or M-24-10, NO authoritative document declares this guide withdrawn. The correction is a DE FACTO withdrawal inferred from two verified facts - the mandate was revoked verbatim (quoted above) and the sole canonical host (itvmo.gsa.gov/genai) returned 200 through 2025-01-09 and has 404'd continuously since 2025-01-24. OMB M-25-22 rescinds only M-24-18 and never names this guide. What is certain is that “in_force” was unsupportable; the replacement value is the best available, not a documented repeal."},{"id":"india-dpdp-commencement","targetKind":"instrument-status","targetSlug":"india-dpdpa","targetLabel":"India DPDP Act listed as in force from 2025-01-01","claimBefore":"The catalog listed the DPDP Act 2023 as “in_force” with effectiveDate 2025-01-01.","claimAfter":"The Act is non-self-executing (§1(2)); no provision commenced until MeitY's G.S.R. 843(E) (gazetted 2025-11-14) set a three-phase commencement. Corrected to “partial”, effectiveDate 2025-11-14, with the substantive core not due until ~2027-05-14.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.meity.gov.in/static/uploads/2025/11/53450e6e5dc0bfa85ebd78686cadad39.pdf","quote":"MINISTRY OF ELECTRONICS AND INFORMATION TECHNOLOGY NOTIFICATION New Delhi, the 13th November, 2025 - G.S.R. 846(E) ... 1. Short title and commencement. - (1) These rules may be called the Digital Personal Data Protection Rules, 2025.","authority":"Gazette of India, Extraordinary, Part II-Sec. 3(i) - MeitY commencement notification"},"note":"The prior claim was not merely stale but never true: no provision of the Act was in force on 2025-01-01. A phased-commencement statute cannot be modelled by a single effectiveDate, so the note carries the full three-phase schedule."},{"id":"bletchley-declaration-date","targetKind":"instrument-status","targetSlug":"bletchley-declaration","targetLabel":"Bletchley Declaration adoption date","claimBefore":"The catalog recorded the Bletchley Declaration as adopted/effective 2023-11-02, and a canonical-facts pin asserted that date as verified (labelled \"Bletchley Park summit\").","claimAfter":"The declaration was published and signed on 1 November 2023, the summit’s first day; adoptedDate and effectiveDate corrected to 2023-11-01. Status “in_force” affirmed (the GOV.UK content API returns an empty withdrawn_notice).","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration","quote":"Published: 1 November 2023 Last updated: 13 February 2025 - See all updates ... 13 February 2025 Updated to include New Zealand as a signatory of the declaration. 1 November 2023 First published.","authority":"GOV.UK (UK Government) - official publication page + machine-readable content API"},"note":"A date correction, not a real-world change - the prior value conflated the DECLARATION with the SUMMIT that hosted it (1-2 Nov), a span named in the document’s own title. Independently re-verified before the change: the GOV.UK content API reports first_published_at 2023-11-01, and 28 countries + the EU signed on day 1. Notably the error was ENCODED IN A PIN asserting it as a canonical fact, so the catalog’s own tests certified it - only a check that re-reads the primary source could find it."},{"id":"seoul-declaration-date","targetKind":"instrument-status","targetSlug":"seoul-declaration","targetLabel":"Seoul Declaration adoption date","claimBefore":"The catalog recorded the Seoul Declaration as adopted/effective 2024-05-22, and a canonical-facts pin asserted that date as verified (labelled \"Seoul AI Summit MoU\").","claimAfter":"The declaration was agreed on 21 May 2024 - the date is embedded in the document's own official title. Corrected to 2024-05-21.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.gov.uk/government/publications/seoul-declaration-for-safe-innovative-and-inclusive-ai-ai-seoul-summit-2024/seoul-declaration-for-safe-innovative-and-inclusive-ai-by-participants-attending-the-leaders-session-ai-seoul-summit-21-may-2024","quote":"Seoul Declaration for safe, innovative and inclusive AI by participants attending the Leaders' Session: AI Seoul Summit, 21 May 2024","authority":"GOV.UK (UK Government) - official publication of the Seoul Declaration text"},"note":"Self-evidencing: the correct date appears in the instrument’s own official title, and the text says participants \"gathered at the AI Seoul Summit on 21st May 2024\". The prior pin labelled its value a summit \"MoU\" - it was pinned to a different day-2 output, not to this declaration. Like the Bletchley error, it was locked in by a pin that made it look verified."},{"id":"eu-pld-corrigendum-date","targetKind":"instrument-status","targetSlug":"eu-product-liability-directive","targetLabel":"EU Product Liability Directive application date","claimBefore":"The catalog recorded effectiveDate 2026-12-09, and the row's notes said the Directive applies to products placed on the market after 9 December 2026.","claimAfter":"A corrigendum (OJ L, 2026/90364, published 2026-05-07) amended Art. 2(1) to read 8 December 2026. Both the effectiveDate and the notes were corrected.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024L2853R(01)","quote":"On page 11, Article 2(1): for: '1. This Directive shall apply to products placed on the market or put into service after 9 December 2026.', read: '1. This Directive shall apply to products placed on the market or put into service after 8 December 2026.'.","authority":"EUR-Lex (Publications Office of the European Union) - official corrigendum, CELEX 32024L2853R(01)"},"note":"A corrigendum-class error: the original enacted text WAS 9 December, so the catalog was right when written and silently became wrong. The row's prose repeated the same wrong date, so both were corrected together - fixing only the field would have left the row contradicting itself. The prior lastReviewedAt (2026-06-21) postdated the corrigendum by ~6 weeks."},{"id":"dfars-part-204-deviation","targetKind":"instrument-notes","targetSlug":"dfars-252-204","targetLabel":"DFARS Part 204 operative location","claimBefore":"The catalog cited the codified DFARS Subpart 204.73 text (acquisition.gov) as the operative location of the clause regime.","claimAfter":"Status “in_force” is AFFIRMED and unchanged - 252.204-7012 still binds. Recorded as location drift only: DoD Class Deviation 2026-O0043 (effective 2026-02-17) directs contracting officers to use an attached Part 204 in lieu of the codified 48 CFR ch. 2 text.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.acq.osd.mil/dpap/dars/classdev/DFARS_RFO/Part-204/2026-O0043_TAB_A_Deviation_Memo_Part_204.pdf","quote":"Effective February 17, 2026, contracting officers shall use- [...] The attached DFARS Part 204, Administrative and Information Matters, in lieu of the text codified at 48 CFR chapter 2 [...] This class deviation remains in effect until rescinded or incorporated into the FAR, DFARS, and DFARS PGI.","authority":"U.S. Department of Defense, Defense Pricing and Contracting (DPC/DARS) - Class Deviation 2026-O0043"},"note":"Recorded to show the watch does NOT only fire on errors: the drift check proposed “superseded”, and re-reading the deviation memo DISPROVED that - the obligations still bind, only their operative location moved. The status was deliberately left alone and only the citation record corrected."},{"id":"eu-pld-entry-into-force","targetKind":"instrument-notes","targetSlug":"eu-product-liability-directive","targetLabel":"EU Product Liability Directive entry into force","claimBefore":"The catalog's notes said the revised Product Liability Directive was \"in force 18 Nov 2024\" - the date the directive was PUBLISHED in the Official Journal.","claimAfter":"18 Nov 2024 is the OJ publication date; Art. 21 sets entry into force at the twentieth day following publication, i.e. 8 December 2024. The notes now distinguish publication (18 Nov 2024), entry into force (8 Dec 2024), the Art. 2(1) application trigger (products placed on the market after 8 Dec 2026, as corrected) and the Art. 22 transposition deadline (9 Dec 2026).","surfacedBy":"cross-article-consistency","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng","quote":"This Directive shall enter into force on the twentieth day following that of its publication in the Official Journal of the European Union.","authority":"EUR-Lex (Publications Office of the European Union) - Directive (EU) 2024/2853, Art. 21"},"note":"THE ARTICLE WAS RIGHT AND THE CATALOG ROW WAS WRONG - the reverse of the usual direction. The cross-article consistency re-run flagged the `redress` article for saying the PLD \"entered into force on 8 December 2024\", because it disagreed with this row. Re-verification against the directive proved the ARTICLE correct and the ROW wrong: the row had recorded the OJ publication date as the entry-into-force date. A second flag against the same sentence (its \"transpose by 9 December 2026\") was also DISPROVED: the Art. 22 transposition deadline is 9 Dec 2026 and the corrigendum never touched it - it moved only the Art. 2(1) trigger to \"after 8 December 2026\", which is exactly what makes covered products begin on 9 Dec 2026. Both dates are correct and distinct; conflating them would have corrupted a correct article. Recorded as a caution: a consistency check reports a DISAGREEMENT, and which side is wrong is a separate question that must be answered from the primary source."},{"id":"eu-aia-adopted-date","targetKind":"instrument-status","targetSlug":"eu-ai-act","targetLabel":"EU AI Act adoption date","claimBefore":"The catalog recorded the EU AI Act as adopted 2024-07-12, and a canonical-facts pin asserted that date as verified with the rationale \"Official Journal publication\".","claimAfter":"2024-07-12 is the Official Journal PUBLICATION date. Regulation (EU) 2024/1689 was adopted/signed on 13 June 2024 (\"Done at Brussels, 13 June 2024\"); adoptedDate corrected to 2024-06-13. effectiveDate (2024-08-01) is unaffected and verified: Art. 113 sets entry into force at the twentieth day following OJ publication.","surfacedBy":"primary-source-reverification","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32024R1689","quote":"REGULATION (EU) 2024/1689 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 13 June 2024 ... Done at Brussels, 13 June 2024.","authority":"EUR-Lex (Publications Office of the European Union) - CELEX 32024R1689; the same record lists 13/06/2024 as both \"Date of document\" and \"Date of signature\""},"note":"THE FLAGSHIP ROW, AND THE CATALOG'S OWN TESTS CERTIFIED THE ERROR. The pin's rationale was the tell: it read \"(Official Journal publication)\" under a field named adoptedDate - i.e. it named the very event that is NOT adoption. The error also broke the catalog's own convention, under which every other row dates adoption to the signature event (cf. EU-PLD-2024: adopted 2024-10-23, published 2024-11-18). Same conflation class as the Bletchley and Seoul date errors corrected the same day: two distinct legal events collapsed into one field. Note 2024-07-12 is a real and load-bearing date - it is what effectiveDate derives from - it was simply recorded in the wrong field."},{"id":"bletchley-frontier-anchor","targetKind":"instrument-cell","targetSlug":"bletchley-declaration","targetLabel":"Where the Bletchley Declaration defines frontier AI","claimBefore":"The catalog's foundation_models coverage cell anchored its verdict to \"Declaration §1-2 (frontier AI defined as the subject)\".","claimAfter":"The declaration's §1-2 address AI generally and never mention frontier AI; the definition sits in §4 (\"Particular safety risks arise at the 'frontier' of AI, understood as being those highly capable general-purpose AI models\"). The cell was re-anchored to §4. The `governs` verdict is unchanged and correct - only the provision anchor was wrong.","surfacedBy":"adversarial-panel","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023","quote":"Particular safety risks arise at the 'frontier' of AI, understood as being those highly capable general-purpose AI models, including foundation models, that could perform a wide variety of tasks.","authority":"GOV.UK (UK Government) - the authoritative text of the Bletchley Declaration"},"note":"FOUND BY WRITING THE ARTICLE. The AI-drafted abstract repeated the cell's bad anchor as an affirmative claim about the source text (\"the declaration names frontier AI in its opening paragraphs\"), which the adversarial panel could then check against the primary source - and it refuted it verbatim. A mis-anchored citation sitting quietly in a coverage cell is nearly invisible; the same error restated as prose is falsifiable. The prose was held and the CELL corrected, since fixing only the prose would have left the underlying cell wrong. Verified independently before the change: §1 reads \"Artificial Intelligence (AI) presents enormous global opportunities...\" and §2 \"AI systems are already deployed across many domains...\"; neither names frontier AI."},{"id":"bletchley-shared-approach","targetKind":"article-prose","targetSlug":"bletchley-declaration","targetLabel":"Whether the Bletchley Declaration commits signatories to a shared approach","claimBefore":"A drafted abstract described the declaration's ask as \"a shared, risk-based approach across signatories\".","claimAfter":"The declaration commits signatories to \"respective\" risk-based policies and expressly recognises that approaches may differ; the prose was corrected to say so before publication.","surfacedBy":"adversarial-panel","status":"corrected","date":"2026-07-17","externalSource":{"url":"https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023","quote":"building respective risk-based policies across our countries to ensure safety in light of such risks, collaborating as appropriate while recognising our approaches may differ based on national circumstances.","authority":"GOV.UK (UK Government) - the authoritative text of the Bletchley Declaration, §6"},"note":"CAUGHT PRE-PUBLICATION, recorded for symmetry with the cell error above. The draft INVERTED the source: the declaration says \"respective ... approaches may differ\", which is the opposite of a shared approach - a substantive misreading of a multilateral instrument's central compromise, not a wording nit. It never reached publication; the ledger records it because the same sentence's other half (the mis-anchored frontier claim) traced to a catalog cell that HAD been published wrong."}]}