{"entities":[{"id":"openai","name":"OpenAI"},{"id":"anthropic","name":"Anthropic"},{"id":"xai","name":"xAI"},{"id":"rsp-v3-0","name":"Anthropic Responsible Scaling Policy — v3.0"},{"id":"rsp-v3-3","name":"Anthropic Responsible Scaling Policy — v3.3"},{"id":"fsf-v3-0","name":"Google DeepMind Frontier Safety Framework — v3.0"},{"id":"durk-kingma","name":"Durk Kingma"},{"id":"guodong-zhang","name":"Guodong Zhang"},{"id":"adam-paper","name":"Adam: A Method for Stochastic Optimization"},{"id":"rlhf","name":"Reinforcement learning from human feedback"},{"id":"ashish-vaswani","name":"Ashish Vaswani"},{"id":"chatglm-report","name":"ChatGLM research history and techniques"},{"id":"meta-faf-v1-1","name":"Meta Frontier AI Framework — Version 1.1"},{"id":"meta-aasf-v2-0","name":"Meta Advanced AI Scaling Framework — Version 2"},{"id":"california-sb-53","name":"California SB 53 (2025): Transparency in Frontier Artificial Intelligence Act, as chaptered"},{"id":"anthropic-frontier-compliance-framework","name":"Anthropic Frontier Compliance Framework"},{"id":"openai-inc","name":"OpenAI, Inc. (OpenAI Foundation)"},{"id":"openai-gp-llc","name":"OpenAI GP LLC"},{"id":"frontier-ai-safety-commitments-2024","name":"Frontier AI Safety Commitments, AI Seoul Summit 2024"},{"id":"turing-1950-computing-machinery","name":"Computing Machinery and Intelligence"},{"id":"extropians-1998-09-24-hanson","name":"AI big wins (was: Punctuated Equilibrium Theory)"},{"id":"seed-ai","name":"Seed AI"},{"id":"fix-problems-as-they-arise","name":"Problems can be fixed as they arise"},{"id":"sl4-2005-07-25-yudkowsky-0507-11765","name":"Re: 6 points about Coherent Extrapolated Volition"},{"id":"eu-digital-omnibus-ai-2026","name":"Regulation (EU) 2026/1744 (Digital Omnibus on AI), as published in the Official Journal"},{"id":"machines-cheaper-at-every-task","name":"Machines cheaper than labour at every task"},{"id":"task-based-exposure-measurement","name":"Exposure to automation measured from task content"},{"id":"recent-economics-engagement-with-ai","name":"Recent entry of machine intelligence into mainstream economics"},{"id":"acemoglu-restrepo-2016-race-between-machine-and-man","name":"The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment"},{"id":"korinek-stiglitz-2017-ai-income-distribution","name":"Artificial Intelligence and Its Implications for Income Distribution and Unemployment"},{"id":"mansfield-1966-technological-change","name":"Technological Change: Measurement, Determinants, and Diffusion"},{"id":"leontief-duchin-1984-impacts-of-automation","name":"The Impacts of Automation on Employment, 1963–2000. Final Report"},{"id":"leontief-1952-machines-and-man","name":"Machines and Man"},{"id":"pajarinen-rouvinen-2014-computerization-finnish-employment","name":"Computerization Threatens One Third of Finnish Employment"},{"id":"ai-existential-risk-in-economic-analysis","name":"Existential risk from AI as a matter for economic analysis"},{"id":"ai-developer-incentives-and-safety","name":"AI developers’ incentives matter for safety"},{"id":"trammell-korinek-2026-growth-under-transformative-ai-nber-revision","name":"Economic Growth under Transformative AI"},{"id":"barret-zoph","name":"Barret Zoph"},{"id":"essential-ai","name":"Essential AI"},{"id":"meta-superintelligence-labs","name":"Meta Superintelligence Labs"},{"id":"pei-sun","name":"Pei Sun"},{"id":"daron-acemoglu","name":"Daron Acemoglu"},{"id":"robin-hanson","name":"Robin Hanson"},{"id":"wassily-leontief","name":"Wassily Leontief"},{"id":"tc260-asgf-v3-0","name":"AI Safety Governance Framework 3.0 (人工智能安全治理框架3.0)"},{"id":"us-bis-gp10-guidance-2025","name":"Guidance on Application of General Prohibition 10 (GP10) to People’s Republic of China (PRC) Advanced-Computing Integrated Circuits (ICs)"},{"id":"lumen-orbit-wp-2024","name":"Why we should train AI in space"},{"id":"oc7-radiation-effects","name":"Radiation effects on computing hardware in orbit"},{"id":"sk-hynix","name":"SK hynix"},{"id":"us-ferc","name":"Federal Energy Regulatory Commission"},{"id":"bwc-msp-2025-wp27","name":"Statement by the United Kingdom of Great Britain & Northern Ireland (BWC/MSP/2025/WP.27)"},{"id":"us-hhs-synthetic-dsdna-guidance-2010","name":"Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA"},{"id":"us-commerce-caisi-bulletin-2025","name":"Department of Commerce bulletin of 4 June 2025 on the Center for AI Standards and Innovation"},{"id":"claude-3-7-sonnet-system-card","name":"Claude 3.7 Sonnet System Card"}],"claims":[{"id":"guodong-zhang-xai-team-2023","statement":"Guodong Zhang appeared on xAI’s July 2023 launch team roster.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Contemporaneous secondary report; original archive unavailable. Roster membership does not establish a co-founder title, start date or current role.","evidenceIds":["evidence-guodong-zhang-xai-team-2023"],"attributes":{"role":"Launch team member"}},{"id":"lineage-kingma-adam","statement":"Adam: A Method for Stochastic Optimization credits Diederik P. Kingma (Durk Kingma) as a co-author.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"One credited co-author, not sole authorship, leadership, employment or priority for the method.","evidenceIds":["evidence-lineage-kingma-adam","evidence-lineage-kingma-identity"]},{"id":"cn-chatglm-rlhf","statement":"The ChatGLM report describes instruction tuning followed by SFT and RLHF, with ChatGPT explicitly named as a motivation.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Retrospective authors’ account in §1. It does not establish model copying, a corporate relationship or independently verified training performance.","evidenceIds":["evidence-cn-chatglm-rlhf"]},{"id":"supersedes-meta-aasf-v2-0","statement":"Meta’s Advanced AI Scaling Framework, Version 2, replaces the Frontier AI Framework, Version 1.1, in the publisher’s documented framework sequence.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"This records replacement in the publisher’s documented framework sequence. It does not retire historical claims or prove implementation. Meta calls Version 2 the second iteration of the same framework under a new name.","evidenceIds":["evidence-supersedes-meta-aasf-v2-0","evidence-supersedes-meta-aasf-v2-0-2"]},{"id":"sb-anthropic-fcf-published","statement":"Anthropic announced on 19 December 2025 that its Frontier Compliance Framework was available, stating that the framework will serve as its compliance framework for SB 53 and other regulatory requirements while its Responsible Scaling Policy remains its voluntary safety policy.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An announcement by Anthropic about its own document; the framework’s own text, version and dates were not inspected in this pass, so no record links the framework itself to SB 53. The stated purpose is forward-looking (“will serve”). It is not evidence that the framework satisfies the statute or that Anthropic is a large frontier developer under it. Anthropic makes this statement in its announcement, not in the FCF itself, which was not retrieved. A designation is not evidence that the FCF satisfies § 22757.12 or that Anthropic complies. The same post says Anthropic endorsed SB 53 (its stated position, not recorded here). For the RSP side of the split, see the record of RSP v3.0’s reference to SB 53.","evidenceIds":["evidence-sb-anthropic-fcf-published-1","evidence-sb-anthropic-fcf-published-2","evidence-sb-anthropic-fcf-published-3"]},{"id":"cs-openai-990-2019-inc-controls-gp-llc","statement":"OpenAI, Inc.’s Form 990 for tax year 2019 lists OpenAI GP LLC as a disregarded entity whose direct controlling entity is OpenAI Inc.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A self-reported tax schedule. “Direct controlling entity” is the form’s term and is not interpreted here under corporate law. The schedule gives CA as the legal domicile code of OpenAI GP LLC and OpenAI LP, while a 2023 court filing describes both as Delaware entities; neither is recorded as a legal form.","evidenceIds":["evidence-cs-openai-990-2019-inc-controls-gp-llc-1"],"attributes":{"control":{"right":"controls","instrument":"Form 990, tax year 2019, Schedule R Part I: direct controlling entity of a disregarded entity","stage":"effective"}}},{"id":"prov-fasc-2024-intolerable-thresholds","statement":"Commitment II states that the organisations will set out thresholds at which severe risks posed by a model or system, unless adequately mitigated, would be deemed intolerable.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the text says, not evidence of any threshold an organisation set.","evidenceIds":["evidence-prov-fasc-2024-intolerable-thresholds-1"],"attributes":{"provision":{"kind":"threshold","key":"intolerable-risk-thresholds"}}},{"id":"ts-a2-turing-1950-context","statement":"Turing’s 1950 paper proposes programming a child-like machine and educating it, and says a machine could help make up its own programmes and modify them to achieve a purpose more effectively.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Education and self-programming toward an unspecified purpose; not a seed designed to redesign its own intelligence.","evidenceIds":["evidence-ts-a2-turing-1950-context","evidence-ts-a2-turing-1950-context-2"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"The child programme is educated and selected by an experimenter, and a self-programming machine is given no starting architecture for growth in intelligence."}}},{"id":"yk-y14-sl4-2005-07-25-yudkowsky-0507-11765-b","statement":"The SL4 message “Re: 6 points about Coherent Extrapolated Volition” (2005-07-25) bears on Y14. In a 2005 thread on Coherent Extrapolated Volition, he lists what Creating Friendly AI (2001) did and did not propose; the next paragraph says he first realised in 2002 that almost every possible AI design would kill him and that until then 'nothing you do is serious safetywise'.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Premise addressed: We get to learn from failures with powerful AI and try again, as with other technologies. Retrospective: the inner quotation is his 2005 paraphrase of CFAI and was not found verbatim in the cached CFAI reprint. The rejection is in the following paragraph and is implicit (it does not name this plan). The earlier view is about building AI and Friendliness in parallel, not about learning from failures of powerful AI, so the fit to the premise is partial; needs review. Proposed as an earlier view; recorded as related because no later text located so far explicitly rejects this earlier statement. His 2005 description of the plan in Creating Friendly AI (2001), rejected in the same message; the inner quotation may paraphrase rather than quote CFAI. Omitted tail: \"based on the stated assumption that the AI part was ten times as much work as the F part\". Cross-family check: the plan combined AI and Friendliness development; it does not endorse learning from failures of powerful AI, and the rejection is implicit.","evidenceIds":["evidence-yk-t14-s3"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"A later description of a plan to build AI and Friendliness in parallel; it does not address learning from failures with powerful AI.","retrospective":{"from":"2001","until":"2002","note":"Reports the plan of Creating Friendly AI (2001), which he says he gave up after realising in 2002 that almost every AI design would kill him."}}}},{"id":"omni-art56-adequacy-assessment","statement":"Article 1, point (21), of Regulation (EU) 2026/1744 replaces Article 56(6) of the AI Act so that the Commission, taking utmost account of the opinion of the Board, assesses whether the codes of practice cover the obligations in Articles 53 and 55, and publishes its assessment of their adequacy.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The amending text. The replaced paragraph in the 2024 text gave the assessment to the AI Office and the Board (separate record). The AI Office is a function of the Commission, and the 2024 Article 3 states that references to the AI Office are construed as references to the Commission. The amendment also removes, in the inspected text of the new paragraph, the sentence on approving a code by implementing act. The recital quoted as context describes codes as having limited legal effect and not granting a presumption of conformity; it is a recital, not an operative provision, and Selekoa does not resolve how it relates to Articles 53(4) and 55(2). Recording an amendment is not evidence that any code was assessed under it.","evidenceIds":["evidence-omni-art56-adequacy-assessment-1","evidence-omni-art56-adequacy-assessment-2","evidence-omni-art56-adequacy-assessment-3"],"attributes":{"provision":{"kind":"decision_authority","key":"omni-code-adequacy-assessment","parties":[{"role":"decides","actor":"Commission, taking utmost account of the opinion of the Board"}]}}},{"id":"ec-m09-mansfield-1966-technological-change-b","statement":"Mansfield, “Technological Change: Measurement, Determinants, and Diffusion” (1966-02) bears on M09. Speaker: Edwin Mansfield, University of Pennsylvania (economist). Mansfield reviews output-per-man-hour evidence (Clague and Greenberg, Kendrick and Sato, CEA) for acceleration.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The passage reports a view the author does not state as their own. The raw djvu text contains this page twice (duplicate scan). Matches the cached .txt. The sentence continues: \"and it is equally possible to find statements asserting the opposite\".","evidenceIds":["evidence-ec-m09-04"],"attributes":{"thought":{"roles":["context"],"register":"relay","missing":"Shows that economists were debating whether automation-era technical change was faster, but is not a statement about machine intelligence as a research topic.","tier":{"value":1,"basis":"Economist author (Edwin Mansfield; established record as an economist)."},"conditions":"'and it is equally possible to find statements asserting the opposite'. He quotes Diebold and DuBridge (non-economists) for the opposite view."}}},{"id":"ec-m05-leontief-1952-machines-and-man-b","statement":"Leontief, “Machines and Man” (1952-09) bears on M05. The standfirst of an article by the economist Wassily Leontief on automatic control systems asks whether they will cause widespread technological unemployment.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The passage reports a view the author does not state as their own. Only the publisher's standfirst was retrieved; the article text was not (paywall). The standfirst may be editorial wording rather than Leontief's, and the byline does not authenticate it as his. Dated bibliographic evidence that an economist's article addressed machines and employment in 1952. Tier 1 since 2026-10-11 (magazine article by an economist); the standfirst stays reported voice and cannot hold earliest located.","evidenceIds":["evidence-ec-m05-00"],"attributes":{"thought":{"roles":["context"],"register":"relay","missing":"Poses the question only; the article text, where Leontief answers it, was not retrieved.","tier":{"value":1,"basis":"Article by an economist in a magazine: tier 1 under the 2026-10-11 rule (D2 names this piece). Moved from tier 2. Only the standfirst was retrieved, and it may be editorial wording."}}}},{"id":"ec-m07-pajarinen-rouvinen-2014-computerization-finnish-employment-b","statement":"Pajarinen and Rouvinen, “Computerization Threatens One Third of Finnish Employment” (2014-01-13) bears on M07. The authors caution that susceptibility is not a forecast of job loss.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"ETLA Brief 22, 13 January 2014; abstract only.","evidenceIds":["evidence-ec-m07-09"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"States the limits of the exposure measure rather than a measurement.","tier":{"value":1,"basis":"Research brief by two economists at ETLA, an economics research institute; a short (6-page) research publication with its own estimates. Tier 1 by maintainer decision of 2026-10-12."}}}},{"id":"ec-e03-korinek-stiglitz-2017-ai-income-distribution-a","statement":"Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates E03. The authors discuss whether humans can keep property rights over AI entities. Drawing on Korinek (2017), the authors describe a Malthusian race between humans and autonomous AI entities.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"(First passage) The authors' own reasoning: loss of control framed through property rights. Speculative; in a footnote. (Second passage) A model scenario that assumes AI entities operate autonomously; it describes human decline through factor-price competition, not violent extinction.","evidenceIds":["evidence-ec-e03-05","evidence-ec-e03-06"],"attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","qualifier":"A model scenario under stated conditions, not the author’s own claim.","tier":{"value":1,"basis":"NBER working paper (core venue)."},"conditions":"(E03-05) If the intelligence gap between humans and super-intelligent AI entities grows too large. (E03-06) Humans do not or cannot assert ownership over AI entities; AI reproduction technology improves exponentially."}}},{"id":"ec-e04-trammell-korinek-2026-growth-under-transformative-ai-nber-revision-b","statement":"Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) bears on E04. Closing sentence of the dangerous-AI subsection.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Wording checked against the April 2026 revision of NBER w31815 (first issued October 2023); whether this wording appears in the October 2023 version was not checked, so it is attested only from April 2026.","evidenceIds":["evidence-ec-e04-14"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Lists safety research and governance frameworks as policy responses; it does not analyse developers' incentives, competition or governance structures.","attested":{"date":"2026-04","note":"Revision date on the cover of the retrieved PDF; the October 2023 wording is recorded under source source-trammell-korinek-2023-growth-under-transformative-ai-nber."},"tier":{"value":1,"basis":"NBER working paper (core venue); Korinek is an economist, Trammell an economist (Oxford, GPI)."}}}},{"id":"lg-goo-vaswani-essential-ai-ceo-2023","statement":"Essential AI's press release of 12 December 2023 identifies Ashish Vaswani as CEO of Essential AI.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A dated snapshot; it does not establish a start date or continuing role.","evidenceIds":["lg-ev-goo-vaswani-essential-ai-ceo-2023"],"attributes":{"role":"CEO"}},{"id":"lg-mx-sun-msl-2025","statement":"Mark Zuckerberg's memo of 30 June 2025, as published by CNBC, names Pei Sun, described as having worked on Gemini post-training, coding and reasoning at Google DeepMind, among new MSL team members.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting: the memo text as published by CNBC. Prior work is Meta's characterisation.","evidenceIds":["lg-ev-mx-sun-msl-2025"],"attributes":{"role":"Team member, Meta Superintelligence Labs"}},{"id":"lg-new-zoph-openai-return-2026","statement":"Fidji Simo posted on X on 15 January 2026 (UTC) that Barret Zoph was rejoining OpenAI and would report to her.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An announced return; the post does not state a title or start date.","evidenceIds":["lg-ev-new-zoph-openai-return-2026"],"attributes":{"activityStage":"announced","role":"not stated (rejoining OpenAI, reporting to Fidji Simo)"}},{"id":"cmp-fsf-v3-0-d6-discretion","statement":"FSF · v3.0 states a provision on the comparison question “Other exceptions or discretion”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Risk acceptance criteria are introduced with 'if, for example', so the listed conditions are examples rather than an exhaustive test. The closing Note (PDF p. 7) says assessments 'will often involve some level of subjective analysis'. What the edition states, quoted from the cited expression. It is not evidence of implementation, compliance or effectiveness, and a shared question does not make two editions’ provisions equivalent.","evidenceIds":["ev-cmp-fsf-v3-0-d6-discretion-1"],"attributes":{"provision":{"kind":"exception","key":"comparison-d6-discretion","facet":{"id":"D6.discretion","status":"stated"}}}},{"id":"ap-daron-acemoglu-acemoglu-restrepo-2016-race-between-machine-and-man","statement":"“The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) credits Daron Acemoglu as one of its 2 authors.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Credited co-authorship; it does not allocate parts of the text between authors or imply agreement with every passage.","evidenceIds":["evidence-ap-daron-acemoglu-acemoglu-restrepo-2016-race-between-machine-and-man"]},{"id":"ap-robin-hanson-extropians-1998-09-24-hanson","statement":"The mailing-list message “AI big wins (was: Punctuated Equilibrium Theory)” (1998-09-24) is sent under the name Robin Hanson.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A sender name on a mailing-list message as archived; it does not establish identity beyond the credited name, and quoted text by others inside the message is not the sender’s.","evidenceIds":["evidence-ap-robin-hanson-extropians-1998-09-24-hanson"]},{"id":"ap-wassily-leontief-leontief-duchin-1984-impacts-of-automation","statement":"“The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) credits Wassily Leontief as one of its 2 authors.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The title page names seven further contributors on a separate “with” line. Credited co-authorship; it does not allocate parts of the text between authors or imply agreement with every passage.","evidenceIds":["evidence-ap-wassily-leontief-leontief-duchin-1984-impacts-of-automation"]},{"id":"cmp-rsp-v3-0-d1-persuasion","statement":"No passage meeting the comparison question “Persuasion or manipulation” was found in RSP · v3.0 within the inspected scope.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: the comparison question asks for the following. The edition names persuasion or manipulation of people as a tracked domain. The inspected scope of RSP · v3.0 was: Whole text layer, PDF pp. 1–19 (all sections, appendices and changelog read). Method: the edition was read in full and its text layer searched case-insensitively for persua, manipul, deceiv, influence, propaganda, disinform. Hits examined and not counted: No hits for persua, influence, propaganda or disinform. “manipulate”/“manipulation” and “deceive”/“deception” (PDF pp. 7 and 9) occur in the high-stakes sabotage opportunities and Automated R&D rows: AI systems manipulating how successor systems are trained or how testing works, evidence that models do not tend to deceive or manipulate users, and AI capabilities for deceiving and obscuring evidence. These describe misalignment and sabotage concerns and the evidence a safety case should include, not persuasion or manipulation of people as a tracked domain. Recorded as not found; a reader may treat the propensity to manipulate users as adjacent. The conclusion is bounded by that scope and method.","editorialState":"seed","qualification":"A recorded absence within the stated scope: not a claim that the provision was deleted, and not a statement about practice or intent.","evidenceIds":["ev-cmp-rsp-v3-0-d1-persuasion"],"attributes":{"provision":{"kind":"risk_domain","key":"comparison-d1-persuasion","facet":{"id":"D1.persuasion"}}}},{"id":"cmp-rsp-v3-3-d1-autonomy","statement":"RSP · v3.3 states a provision on the comparison question “Autonomy, self-improvement or AI research and development”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Named as a threshold row on automated R&D in key domains; the edition says its evaluations will for now focus on AI R&D. Autonomous, goal-directed operation appears as part of the “Misaligned AI systems in high-stakes settings” row (see “Other named domains”). No separate threshold for autonomous replication is set. What the edition states, paraphrased from the cited expression with its page. It is not evidence of implementation, compliance or effectiveness, and a shared question does not make two editions’ provisions equivalent.","evidenceIds":["ev-cmp-rsp-v3-3-d1-autonomy-1"],"attributes":{"provision":{"kind":"risk_domain","key":"comparison-d1-autonomy","facet":{"id":"D1.autonomy","status":"stated"}}}},{"id":"c2-governance-tc260-asgf-v3-agent-risk-framework","statement":"The AI Safety Governance Framework 3.0 proposes establishing a risk management framework for AI agents (智能体) that analyses their safety risks and proposes prevention measures; it sets one out in its Annex 2.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What version 3.0 says. 智能体 is rendered as AI agents. Annex 2 (智能体风险管理框架) is listed in the table of contents; its content is not recorded here.","evidenceIds":["c2-ev-governance-tc260-asgf-v3-agent-risk-framework-1"],"attributes":{"provision":{"kind":"required_action","key":"agent-risk-management"}}},{"id":"cmp-meta-faf-v1-1-d1-autonomy","statement":"Meta FAF · v1.1 states a provision on the comparison question “Autonomy, self-improvement or AI research and development”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Named as an enabling capability within the cyber domain, not as a domain with thresholds of its own. The same table names autonomous execution of multi-step tasks, and §4.2 (PDF p. 17) takes account of automated AI research and development that enhances capabilities. No autonomous replication threshold is named. What the edition states, paraphrased from the cited expression with its page. It is not evidence of implementation, compliance or effectiveness, and a shared question does not make two editions’ provisions equivalent.","evidenceIds":["ev-cmp-meta-faf-v1-1-d1-autonomy-1"],"attributes":{"provision":{"kind":"risk_domain","key":"comparison-d1-autonomy","facet":{"id":"D1.autonomy","status":"stated"}}}},{"id":"fl-link-rsp-v3-3-sb53","statement":"Anthropic’s Responsible Scaling Policy v3.3 keeps the footnote saying that where laws such as California SB 53 define “catastrophic risk” or similar terms with specific thresholds, Anthropic addresses those requirements in separate compliance frameworks.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The edition’s own statement about SB 53, quoted from it. It names SB 53 and points to separate compliance frameworks; it does not say that the RSP is a framework under the Act. It is not evidence that the edition satisfies the statute, that its publisher is a frontier developer under it, or of implementation or compliance. The record on v3.0 notes the same footnote in the later editions; this record gives v3.3 its own link, as the dataset does for the White House and Seoul links.","evidenceIds":["evidence-fl-link-rsp-v3-3-sb53-1"],"attributes":{"commitment":{"link":"names"}}},{"id":"cpo-ch-bis-gp10-guidance-ascend-list","statement":"The BIS guidance of 13 May 2025 lists the Huawei Ascend 910B, 910C and 910D as illustrative PRC 3A090 ICs subject to the presumption that General Prohibition 10 restrictions apply, and states that the list is not exhaustive.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The guidance prints the third chip as \"Huawei Ascent 910D\"; the reading \"Ascend\" follows the summary’s \"Huawei Ascend chips\" and the series naming, and is Selekoa’s. The list states a presumption, not a finding about any transaction.","evidenceIds":["evidence-cpo-ch-bis-gp10-guidance-ascend-list-1","evidence-cpo-ch-bis-gp10-guidance-ascend-list-2"],"attributes":{"provision":{"kind":"statement","key":"ch-bis-gp10-guidance-ascend-list"}}},{"id":"cpo-orb-oc7-lumen-2024","statement":"The Lumen Orbit white paper (September 2024) states that shielding is required against radiation-induced latch-up, transients and total-dose effects in sensitive components, while arguing that shielding mass per unit of compute falls with container size.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Earliest located among the texts found by the pack’s searches of 2026-10-28 and the cross-family check’s bounded search of 2026-10-29/30; not a claim that no earlier text exists. From a proponent, which treats radiation as a manageable design cost. It held the role in the pack, lost it at part 1 of the admission to a 1989 paper on space radiation and electronic microcircuits, and holds it again since the maintainer narrowed the criterion on 2026-10-30 to texts that discuss computing placed in orbit for this purpose (a correction recorded on the collection page).","evidenceIds":["evidence-cpo-orb-oc7-lumen-2024-1","evidence-cpo-orb-oc7-lumen-2024-2"],"attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction","attested":{"date":"2024-12-24","note":"The cover prints September 2024 (v1.01). The earliest copy located is an Internet Archive capture of 24 December 2024, which the cross-family check found byte-identical to the copy retrieved for the pack and again at admission."}}}},{"id":"cpo-pw-ferc-susquehanna-isa-rejected","statement":"FERC, by order issued on 1 November 2024 (189 FERC ¶ 61,078, Docket ER24-2172), rejected the amended Interconnection Service Agreement among PJM, Susquehanna Nuclear and PPL Electric Utilities, which would have increased the co-located load at the Susquehanna plant from 300 MW to 480 MW. World Nuclear News reported the rejection on 4 November 2024.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Dated by FERC’s own order (cross-family check, item 44); the press report is context. The order names Susquehanna Nuclear, LLC, not Talen; the report’s “Talen” is the corporate group. The 960 MW of the Cumulus campus is a separate figure, not the rejected load.","evidenceIds":["evidence-cpo-pw-ferc-susquehanna-isa-rejected-1","evidence-cpo-pw-ferc-susquehanna-isa-rejected-2","evidence-cpo-pw-ferc-susquehanna-isa-rejected-3","evidence-cpo-pw-ferc-susquehanna-isa-rejected-4","evidence-cpo-pw-ferc-susquehanna-isa-rejected-5","evidence-cpo-pw-ferc-susquehanna-isa-rejected-6","evidence-cpo-pw-ferc-susquehanna-isa-rejected-7"],"attributes":{"quantity":{"value":480,"unit":"MW co-located load (proposed, up from 300 MW)","bound":"exact","measures":"facility_power"},"power":{"technology":"nuclear","mechanism":"interconnection"}}},{"id":"cpo-ch-skhynix-openai-loi","statement":"SK Group announced on 1 October 2025 that SK hynix had signed a letter of intent with OpenAI to supply High Bandwidth Memory for Stargate data centres and would establish a production system able to meet OpenAI’s HBM demand of up to 900,000 DRAM wafers per month.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A letter of intent; SK’s statement of a production system to be established is a plan, not capacity in place, and no supply volume, price or delivery is shown. SK’s further statement that this demand is \"more than double the current HBM industry capacity\" is SK’s comparison and is not checked here. The SK Telecom MOU on a Korean data centre belongs to the data-centre strand.","evidenceIds":["evidence-cpo-ch-skhynix-openai-loi-1","evidence-cpo-ch-skhynix-openai-loi-2"],"attributes":{"transactionStage":"letter_of_intent","supply":{"product":"memory","units":"HBM; production system for demand of 'up to 900,000 DRAM wafers per month' (plan)"}}},{"id":"ab-scr-hhs-2010-effective-date","statement":"The 2010 notice states that the Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA is effective on 13 October 2010.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"States the date the 2010 guidance gives for its own effect. The provision kind is `date`, added in schema 2.22 for a date a text states for its own application; part 1 recorded it as a threshold, on the dataset’s earlier precedent.","evidenceIds":["evidence-ab-scr-hhs-2010-effective-date-1"],"attributes":{"provision":{"kind":"date","key":"hhs-2010-effective-date"}}},{"id":"ab-gov-bwc-uk-wp27-biotech-ai-threats","statement":"The statement by the United Kingdom issued as document BWC/MSP/2025/WP.27 at the 2025 Meeting of States Parties states that technology is at present outstripping diplomacy and that novel and hybrid threats to collective security have emerged, including using biotechnology and AI as new kinds of weapons of war.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A statement of one State Party's position, in its own general wording. It names no actor, no system and no incident, and it is not a finding that any such use has occurred.","evidenceIds":["evidence-ab-gov-bwc-uk-wp27-biotech-ai-threats-1"],"attributes":{"provision":{"kind":"risk_domain","key":"bwc-uk-statement-biotech-ai-threats"}}},{"id":"ab-gov-us-commerce-caisi-bulletin","statement":"The Department of Commerce bulletin sent on 4 June 2025, carrying a release dated 3 June 2025, states that the Secretary of Commerce announced his plans to reform the agency formerly known as the U.S. AI Safety Institute into the Center for AI Standards and Innovation (CAISI).","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Announced plans, as the bulletin states them. It is not evidence that any reorganisation was completed, and no legal continuity or succession is recorded. The dataset already carries the Center for AI Standards and Innovation as a dated name of the institute from June 2025, on a later NIST editorial note; this record adds the Commerce statement itself, which the earlier pass could not retrieve.","evidenceIds":["evidence-ab-gov-us-commerce-caisi-bulletin-1"],"attributes":{"provision":{"kind":"statement","key":"us-commerce-caisi-bulletin-announcement"}}},{"id":"ab-ev-claude37-asl2-determination","statement":"The Claude 3.7 Sonnet System Card states that ASL-2 safeguards remain appropriate for the model, that its evaluations suggest improved performance in all domains including some uplift in CBRN evaluations, and that on what it observed in recent CBRN testing there is a substantial probability that the next model may require ASL-3 safeguards.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A developer's determination under its own policy, for the model and evaluations the card describes. The stated expectation about a next model is the card's own statement about the future, not a recorded event. The card also states that some specific CBRN thresholds and results are treated as sensitive and not published, so the published basis is incomplete. Only determination-level and caveat sentences are cited; the card's CBRN results section holds technical content, which is excluded.","evidenceIds":["evidence-ab-ev-claude37-asl2-determination-1","evidence-ab-ev-claude37-asl2-determination-2","evidence-ab-ev-claude37-asl2-determination-3","evidence-ab-ev-claude37-asl2-determination-4"],"attributes":{"evaluatedSystem":"Claude 3.7 Sonnet, final release model unless the card states otherwise","evaluation":{"what":"Whether the model reaches the ASL-3 capability threshold for CBRN, with biological risks the primary focus","how":"Automated knowledge and skill-testing evaluations, uplift studies designed to proxy real-world tasks, external expert red teaming and long-form agentic evaluations","result":"ASL-2 safeguards remain appropriate; improved performance in all domains, including some uplift in CBRN evaluations; a substantial probability that the next model may require ASL-3 safeguards","caveats":"The card states that some specific evaluation thresholds and results in the CBRN domain are considered sensitive and are not shared publicly, and that how the model performs on expert-level evaluations may not directly translate to real-world capabilities."}}}],"sources":[{"id":"source-xai-team-2023","title":"Musk announces new AI company that seeks to “understand the universe”","authors":["Benj Edwards"],"publisher":"Ars Technica","publicationDate":"2023-07-12","version":"Contemporaneous report and screenshot of the launch roster. The current x.ai company page has changed; original archive retrieval was unavailable.","canonicalUrl":"https://arstechnica.com/information-technology/2023/07/musk-announces-new-ai-company-that-seeks-to-understand-the-universe/","evidenceUrl":"https://arstechnica.com/information-technology/2023/07/musk-announces-new-ai-company-that-seeks-to-understand-the-universe/","language":"en","archivedUrl":"https://web.archive.org/web/20260211075334/https://arstechnica.com/information-technology/2023/07/musk-announces-new-ai-company-that-seeks-to-understand-the-universe/","archivedAt":"2026-02-11","accessedAt":"2026-09-18","reuse":{"attribution":"Benj Edwards, Musk announces new AI company that seeks to “understand the universe” (2023-07-12).","licence":"No open redistribution licence asserted","note":"Bibliographic metadata and brief attributed paraphrases or excerpts only. Source roles are relative to each claim."}},{"id":"source-rsp-3-0","title":"Responsible Scaling Policy — v3.0","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":"2026-02-24","version":"Responsible Scaling Policy v3.0; effective date 2026-02-24.","versionDate":"2026-02-24","canonicalUrl":"https://anthropic.com/responsible-scaling-policy/rsp-v3-0","evidenceUrl":"https://anthropic.com/responsible-scaling-policy/rsp-v3-0","language":"en","archivedUrl":"https://web.archive.org/web/20260224203046/https://www.anthropic.com/responsible-scaling-policy/rsp-v3-0","archivedAt":"2026-02-24","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v3.0.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-rsp-3-3","title":"Responsible Scaling Policy — v3.3","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Responsible Scaling Policy v3.3; effective date 2026-05-26.","versionDate":"2026-05-26","canonicalUrl":"https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","evidenceUrl":"https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260528170023/https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","archivedAt":"2026-05-28","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v3.3.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-fsf-3-0","title":"Frontier Safety Framework — v3.0","authors":["Google DeepMind"],"publisher":"Google DeepMind","publicationDate":"2025-09-22","version":"Frontier Safety Framework v3.0; publication date 2025-09-22.","versionDate":"2025-09-22","canonicalUrl":"https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3.pdf","evidenceUrl":"https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20250922134959/https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3.pdf","archivedAt":"2025-09-22","accessedAt":"2026-09-18","reuse":{"attribution":"Google DeepMind, Frontier Safety Framework — v3.0.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-kingma-bio","title":"Durk Kingma — Brief Bio","authors":["Diederik P. Kingma"],"publisher":"Durk Kingma","publicationDate":null,"version":"Page inspected 2026-09-19; historical statements are scoped to their source, not treated as current appointments.","canonicalUrl":"https://dpkingma.com/","evidenceUrl":"https://dpkingma.com/","language":"en","archivedUrl":"https://web.archive.org/web/20260807144357/https://dpkingma.com/","archivedAt":"2026-08-07","accessedAt":"2026-09-19","reuse":{"attribution":"Diederik P. Kingma, Durk Kingma — Brief Bio","licence":"No open redistribution licence asserted","note":"Bibliographic metadata and brief attributed paraphrases only; no full source or personal contact details redistributed."}},{"id":"source-adam-paper-v1","title":"Adam: A Method for Stochastic Optimization","authors":["Diederik P. Kingma","Jimmy Ba"],"publisher":"arXiv","publicationDate":"2014-12-22","version":"arXiv:1412.6980v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2014-12-22","canonicalUrl":"https://arxiv.org/abs/1412.6980","evidenceUrl":"https://arxiv.org/pdf/1412.6980v1","language":"en","accessedAt":"2026-09-20","reuse":{"attribution":"Diederik P. Kingma; Jimmy Ba, Adam: A Method for Stochastic Optimization (2014-12-22).","licence":"No open redistribution licence asserted","note":"Bibliographic metadata and brief attributed paraphrases only; no full source, personal contact details or dataset imported."}},{"id":"source-cn-chatglm","title":"ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools","authors":["Team GLM"],"publisher":"Team GLM","publicationDate":"2024-06-18","version":"Source language: English (en). arXiv:2406.12793v1; joint Zhipu AI / Tsinghua University affiliation.","canonicalUrl":"https://arxiv.org/html/2406.12793v1","evidenceUrl":"https://arxiv.org/html/2406.12793v1","language":"en","accessedAt":"2026-09-29","reuse":{"attribution":"Team GLM, ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools (2024-06-18).","licence":"No redistribution licence asserted for this source expression","note":"Bibliographic metadata and brief attributed paraphrases only; model licensing is distinct from source-document reuse."}},{"id":"source-meta-faf-v1-1","title":"Frontier AI Framework — Version 1.1","authors":["Meta"],"publisher":"Meta","publicationDate":"2025-02-03","version":"Frontier AI Framework, Version 1.1, as captured by the Internet Archive on 2025-02-13, the earliest complete capture located. The original address has returned HTTP 500 since April 2026; a revised expression under the same label was captured from 2025-03-28 and is recorded separately.","versionDate":"2025-02-03","canonicalUrl":"https://ai.meta.com/static-resource/meta-frontier-ai-framework/","evidenceUrl":"https://web.archive.org/web/20250213105620/https://ai.meta.com/static-resource/meta-frontier-ai-framework/","language":"en","archivedUrl":"https://web.archive.org/web/20250213105620/https://ai.meta.com/static-resource/meta-frontier-ai-framework/","archivedAt":"2025-02-13","accessedAt":"2026-10-03","reuse":{"attribution":"Meta, Frontier AI Framework, Version 1.1 (2025).","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-meta-aasf-v2","title":"Advanced AI Scaling Framework — Version 2","authors":["Meta"],"publisher":"Meta","publicationDate":"2026-04-08","version":"Advanced AI Scaling Framework, Version 2; its change log dates the version April 7, 2026. The file served at the recorded address is byte-identical to the Internet Archive capture of 2026-04-08.","versionDate":"2026-04-07","canonicalUrl":"https://ai.meta.com/static-resource/Meta_Advanced-AI-Scaling-Framework-v2","evidenceUrl":"https://ai.meta.com/static-resource/Meta_Advanced-AI-Scaling-Framework-v2","language":"en","archivedUrl":"https://web.archive.org/web/20260408163104/https://ai.meta.com/static-resource/Meta_Advanced-AI-Scaling-Framework-v2","archivedAt":"2026-04-08","accessedAt":"2026-10-03","reuse":{"attribution":"Meta, Advanced AI Scaling Framework, Version 2 (2026).","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-meta-aasf-announcement","title":"Scaling How We Build and Test Our Most Advanced AI","authors":["Meta"],"publisher":"Meta","publicationDate":"2026-04-08","version":"Dated blog post; quotations match the live page and the Internet Archive capture of 2026-04-08.","canonicalUrl":"https://ai.meta.com/blog/scaling-how-we-build-test-advanced-ai/","evidenceUrl":"https://ai.meta.com/blog/scaling-how-we-build-test-advanced-ai/","language":"en","archivedUrl":"https://web.archive.org/web/20260408173610/https://ai.meta.com/blog/scaling-how-we-build-test-advanced-ai/","archivedAt":"2026-04-08","accessedAt":"2026-10-03","reuse":{"attribution":"Meta, Scaling How We Build and Test Our Most Advanced AI (2026-04-08).","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-anthropic-fcf-announcement","title":"Sharing our compliance framework for California's Transparency in Frontier AI Act","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":"2025-12-19","version":"Web page dated Dec 19, 2025, inspected live 2026-10-04; quote also matched in the 2025-12-28 capture","canonicalUrl":"https://www.anthropic.com/news/compliance-framework-SB53","evidenceUrl":"https://www.anthropic.com/news/compliance-framework-SB53","language":"en","archivedUrl":"https://web.archive.org/web/20251228124116/https://www.anthropic.com/news/compliance-framework-SB53","archivedAt":"2025-12-28","accessedAt":"2026-10-04","reuse":{"attribution":"Anthropic, Sharing our compliance framework for California's Transparency in Frontier AI Act.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-openai-990-2019-schedule-r","title":"Form 990 Schedule R, tax year 2019: OpenAI Inc (ProPublica render)","authors":["OpenAI Inc (filer)"],"publisher":"OpenAI Inc (filer)","publicationDate":"2020-11-16","version":"ProPublica render of IRS e-file ObjectId 202003219349325305, submission dated 2020-11-16, inspected 2026-10-04; row values read from the render markup. The Internet Archive holds only redirect captures of this render.","canonicalUrl":"https://projects.propublica.org/nonprofits/full_text/202003219349325305/IRS990ScheduleR","evidenceUrl":"https://projects.propublica.org/nonprofits/full_text/202003219349325305/IRS990ScheduleR","language":"en","archivedUrl":"https://web.archive.org/web/20250303135109/https://projects.propublica.org/nonprofits/full_text/202003219349325305/IRS990ScheduleR","archivedAt":"2025-03-03","accessedAt":"2026-10-04","reuse":{"attribution":"OpenAI Inc, Form 990 Schedule R for tax year 2019, rendered by ProPublica Nonprofit Explorer.","licence":"Public tax filing; no redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-fasc-2024","title":"Frontier AI Safety Commitments, AI Seoul Summit 2024","authors":["Department for Science, Innovation and Technology"],"publisher":"Department for Science, Innovation and Technology","publicationDate":"2024-05-21","version":"GOV.UK HTML publication, first published 21 May 2024 and updated 7 February 2025 to add organisations to the list; inspected live 2026-10-04. The Internet Archive capture of 2024-05-21 (11:05 UTC) has the sixteen-name list only, and the commitment text quoted is identical in both. Publisher as printed on the inspected expression: Department for Science, Innovation and Technology (GOV.UK).","versionDate":"2025-02-07","canonicalUrl":"https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024","evidenceUrl":"https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024","language":"en","archivedUrl":"https://web.archive.org/web/20240521110530/https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024","archivedAt":"2024-05-21","accessedAt":"2026-10-04","reuse":{"attribution":"Department for Science, Innovation and Technology, Frontier AI Safety Commitments, AI Seoul Summit 2024 (GOV.UK).","licence":"Open Government Licence v3.0 (GOV.UK content); quoted briefly under the editorial policy","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-turing-1950-amt-b-19","title":"Computing Machinery and Intelligence (offprint, AMT/B/19)","authors":["A. M. Turing"],"publisher":"Mind","publicationDate":"1950-10","version":"Original Mind offprint, scanned by the archive; printed pp. 448–449 read on the page images for this collection (scans 16–17). Offprint held by the Turing Digital Archive, King’s College, Cambridge.","canonicalUrl":"https://turingarchive.kings.cam.ac.uk/publications-lectures-and-talks-amtb/amt-b-19","evidenceUrl":"https://turingarchive.kings.cam.ac.uk/publications-lectures-and-talks-amtb/amt-b-19","language":"en","archivedUrl":"https://web.archive.org/web/20261030045116/https://mediadepot.kings.cam.ac.uk/getmedia/81/be4a3d910d5d58e3911315113c60a09f6293fa896a5e3795998fa6b4fb72b71446336e3c7c92d5e08fd0fb2762b6f663279ba799d7fd562480b4bf63c63e02b9","archivedAt":"2026-10-30","accessedAt":"2026-10-06","reuse":{"attribution":"A. M. Turing, \"Computing Machinery and Intelligence\", Mind 59 (236), October 1950; offprint in the Turing Digital Archive, AMT/B/19.","licence":"No open redistribution licence asserted; the archive hosts the offprint with Oxford University Press’s permission. Page images are not reproduced.","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-extropians-1998-09-24-hanson","title":"AI big wins (was: Punctuated Equilibrium Theory) (Extropians mailing list)","authors":["Robin Hanson"],"publisher":"Extropians mailing list","publicationDate":"1998-09-24","version":"Archive HTML as served, retrieved 2026-10-06 with the hash recorded by the 1996–2022 candidate search; quoted lines excluded when reading the author’s text. Archive at extropians.weidai.com.","canonicalUrl":"http://extropians.weidai.com/extropians.3Q98/2611.html","evidenceUrl":"http://extropians.weidai.com/extropians.3Q98/2611.html","language":"en","archivedUrl":"https://web.archive.org/web/20150428002928/http://extropians.weidai.com/extropians.3Q98/2611.html","archivedAt":"2015-04-28","accessedAt":"2026-10-06","reuse":{"attribution":"Robin Hanson, \"AI big wins (was: Punctuated Equilibrium Theory)\", Extropians mailing list, 1998-09-24.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-sl4-2005-07-25-yudkowsky-0507-11765","title":"Re: 6 points about Coherent Extrapolated Volition (SL4 mailing list)","authors":["Eliezer Yudkowsky"],"publisher":"SL4 mailing list","publicationDate":"2005-07-25","version":"Archive HTML as served, retrieved 2026-10-07 (SHA-256 08cb8b0fc83f702f32882f8b3eba0724135658d8406a8c761c4617d8601412fb); quoted lines excluded when reading the author’s text. The served archive shows later URL rewriting and does not authenticate historical bytes. Archive at sl4.org.","canonicalUrl":"http://sl4.org/archive/0507/11765.html","evidenceUrl":"http://sl4.org/archive/0507/11765.html","language":"en","archivedUrl":"https://web.archive.org/web/20251116042043/http://sl4.org/archive/0507/11765.html","archivedAt":"2025-11-16","accessedAt":"2026-10-07","reuse":{"attribution":"Eliezer Yudkowsky, “Re: 6 points about Coherent Extrapolated Volition”, SL4 mailing list, 2005-07-25.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-eu-digital-omnibus-ai-2026","title":"Regulation (EU) 2026/1744 of the European Parliament and of the Council of 8 July 2026 amending Regulations (EU) 2024/1689, (EU) 2018/1139 and (EU) 2023/1230 as regards the simplification of the implementation of harmonised rules on artificial intelligence (Digital Omnibus on AI)","authors":["European Parliament and Council of the European Union"],"publisher":"Official Journal of the European Union (EUR-Lex)","publicationDate":"2026-07-24","version":"Official Journal L series edition of 24.7.2026, English HTML, retrieved 2026-10-10. Its text was identical to that of the Internet Archive capture of 2026-07-28.","canonicalUrl":"https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_202601744","evidenceUrl":"https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_202601744","language":"en","archivedUrl":"https://web.archive.org/web/20260728193314/https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_202601744","archivedAt":"2026-07-28","accessedAt":"2026-10-10","reuse":{"attribution":"European Parliament and Council of the European Union, Regulation (EU) 2026/1744 (Digital Omnibus on AI), Official Journal L, 24.7.2026.","licence":"Official legal text; no redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-acemoglu-restrepo-2016-race-between-machine-and-man","title":"The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment","authors":["Daron Acemoglu","Pascual Restrepo"],"publisher":"NBER Working Paper 22252","publicationDate":"2016-05","version":"Full text (revised version). Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 441d01202afd56ef8002fc24ffc2beb51191741c0b5accb11d2534620dd616b7. Journal version not retrieved. Acemoglu, D., & Restrepo, P. (2018). The race between man and machine: Implications of technology for growth, factor shares, and employment. American Economic Review, 108(6), 1488–1542. https://doi.org/10.1257/aer.20160696","canonicalUrl":"https://www.nber.org/system/files/working_papers/w22252/w22252.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w22252/w22252.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Daron Acemoglu, Pascual Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment”, NBER Working Paper 22252, 2016-05.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-korinek-stiglitz-2017-ai-income-distribution","title":"Artificial Intelligence and Its Implications for Income Distribution and Unemployment","authors":["Anton Korinek","Joseph E. Stiglitz"],"publisher":"NBER Working Paper 24174","publicationDate":"2017-12","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 aeddf962b5c488cae68b7d48cef964de22379c78e75579c5793f0df71d5e9f40.","canonicalUrl":"https://www.nber.org/system/files/working_papers/w24174/w24174.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w24174/w24174.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Anton Korinek, Joseph E. Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment”, NBER Working Paper 24174, 2017-12.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-mansfield-1966-technological-change","title":"Technological Change: Measurement, Determinants, and Diffusion","authors":["Edwin Mansfield"],"publisher":"In The Employment Impact of Technological Change, Appendix Volume II (U.S. GPO, February 1966)","publicationDate":"1966-02","version":"Internet Archive OCR text (djvu.txt) of a scanned copy, retrieved 2026-10-09 from https://dn710704.ca.archive.org/0/items/ERICED022857/ERICED022857djvu.txt; OCR errors possible; quotes checked only against clean stretches. Current copy; no earlier capture compared. SHA-256 85b1fb59b997e8cbc0f7255076a55f68008955a2b10523b375ac0ae58628572c.","canonicalUrl":"https://archive.org/details/ERIC_ED022857","evidenceUrl":"https://archive.org/details/ERIC_ED022857","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Edwin Mansfield, “Technological Change: Measurement, Determinants, and Diffusion”, In The Employment Impact of Technological Change, Appendix Volume II (U.S. GPO, February 1966), 1966-02.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-leontief-duchin-1984-impacts-of-automation","title":"The Impacts of Automation on Employment, 1963–2000. Final Report","authors":["Wassily Leontief","Faye Duchin"],"publisher":"New York: Institute for Economic Analysis, New York University; National Science Foundation contract PRA-8012844; April 1984, 376 pp.; ERIC ED 241 743","publicationDate":"1984-04","version":"Internet Archive OCR text (djvu.txt) of a scanned copy, retrieved 2026-10-09 from https://dn711508.ca.archive.org/0/items/microIA411534480031/microIA411534480031djvu.txt; OCR errors possible; quotes checked only against clean stretches. Current copy; no earlier capture compared. SHA-256 cd03df677ca1a723f843eccec0fa1415b3227d604bc83c6783a40321b9a785e9. Book version not retrieved; the 1984 report was. Leontief, W., & Duchin, F. (1986). The future impact of automation on workers. Oxford University Press.","canonicalUrl":"https://archive.org/details/micro_IA41153448_0031","evidenceUrl":"https://archive.org/details/micro_IA41153448_0031","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Wassily Leontief, Faye Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report”, New York: Institute for Economic Analysis, New York University; National Science Foundation contract PRA-8012844; April 1984, 376 pp.; ERIC ED 241 743, 1984-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-leontief-1952-machines-and-man","title":"Machines and Man","authors":["Wassily Leontief"],"publisher":"Scientific American 187(3), September 1952, pp. 150–160; doi:10.1038/scientificamerican0952-150","publicationDate":"1952-09","version":"Publisher article page (paywalled): title, standfirst and bibliographic line only; the article text was not retrieved. Current copy retrieved 2026-10-09. SHA-256 f9161d3789c10d798b9d27dad8762bef7874d09449ea5dad786db5ae180ad24f. Article text not retrieved; only the standfirst on the publisher page. Leontief, W. (1952). Machines and man. Scientific American, 187(3), 150–160. https://doi.org/10.1038/scientificamerican0952-150","canonicalUrl":"https://www.scientificamerican.com/article/machines-and-man/","evidenceUrl":"https://www.scientificamerican.com/article/machines-and-man/","language":"en","archivedUrl":"https://web.archive.org/web/20241108120949/https://www.scientificamerican.com/article/machines-and-man/","archivedAt":"2024-11-08","accessedAt":"2026-10-09","reuse":{"attribution":"Wassily Leontief, “Machines and Man”, Scientific American 187(3), September 1952, pp. 150–160; doi:10.1038/scientificamerican0952-150, 1952-09.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-pajarinen-rouvinen-2014-computerization-finnish-employment","title":"Computerization Threatens One Third of Finnish Employment","authors":["Mika Pajarinen","Petri Rouvinen"],"publisher":"ETLA Muistio – ETLA Brief 22 (Helsinki: ETLA), 6 pp., ISSN 2323-2463","publicationDate":"2014-01-13","version":"ETLA publication page with abstract (the PDF was not retrieved). Current copy retrieved directly on 2026-10-09. SHA-256 f14a8103d4eb6603c6092abe47665ed38e408020064249fedf4c0bf5747d7f9e.","canonicalUrl":"https://www.etla.fi/julkaisut/computerization-threatens-finnish-employment/","evidenceUrl":"https://www.etla.fi/julkaisut/computerization-threatens-finnish-employment/","language":"en","archivedUrl":"https://web.archive.org/web/20260311035015/https://www.etla.fi/julkaisut/computerization-threatens-finnish-employment/","archivedAt":"2026-03-11","accessedAt":"2026-10-09","reuse":{"attribution":"Mika Pajarinen, Petri Rouvinen, “Computerization Threatens One Third of Finnish Employment”, ETLA Muistio – ETLA Brief 22 (Helsinki: ETLA), 6 pp., ISSN 2323-2463, 2014-01-13.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-trammell-korinek-2026-growth-under-transformative-ai-nber-revision","title":"Economic Growth under Transformative AI","authors":["Philip Trammell","Anton Korinek"],"publisher":"NBER Working Paper 31815","publicationDate":"2023-10","version":"Current NBER PDF, the April 2026 revision. The October 2023 version was retrieved separately from a Wayback capture of 2023-11-29 (source source-trammell-korinek-2023-growth-under-transformative-ai-nber); their wording differs. SHA-256 06dda75180302720e0843a180c00d04fa99e38db0f5d6d4e06d7f29388a115bb. Version a reader would cite, or full text, not retrieved. Trammell, P., & Korinek, A. (2023). Economic growth under transformative AI (NBER Working Paper No. 31815, original October 2023 version). https://doi.org/10.3386/w31815","canonicalUrl":"https://www.nber.org/system/files/working_papers/w31815/w31815.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w31815/w31815.pdf","language":"en","accessedAt":"2026-10-11","reuse":{"attribution":"Philip Trammell, Anton Korinek, “Economic Growth under Transformative AI”, NBER Working Paper 31815, 2023-10.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-goo-essential-series-a-2023","title":"Essential AI Raises $56.5M Series A to Build the Enterprise Brain (press release)","authors":["Essential AI (press release as republished by HPCwire/AIwire, Off the Wire)"],"publisher":"Essential AI (press release as republished by HPCwire/AIwire, Off the Wire)","publicationDate":"2023-12-12","version":"Dated page, retrieved on 2026-10-15 (SHA-256 dd5d42f4827312caeff44ee3cbddb500d7650a9b8a677ec414dd239067086c5e) through the third-party reader relay r.jina.ai, which renders the publisher’s current page (the publisher refused direct retrieval); not an archive copy.","canonicalUrl":"https://www.hpcwire.com/aiwire/2023/12/12/essential-ai-raises-56-5m-series-a-to-build-the-enterprise-brain/","evidenceUrl":"https://www.hpcwire.com/aiwire/2023/12/12/essential-ai-raises-56-5m-series-a-to-build-the-enterprise-brain/","language":"en","archivedUrl":"https://web.archive.org/web/20261029004627/https://www.hpcwire.com/aiwire/2023/12/12/essential-ai-raises-56-5m-series-a-to-build-the-enterprise-brain/","archivedAt":"2026-10-29","accessedAt":"2026-10-15","reuse":{"attribution":"Essential AI (press release as republished by HPCwire/AIwire, Off the Wire), “Essential AI Raises $56.5M Series A to Build the Enterprise Brain (press release)”, 2023-12-12.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-mx-cnbc-msl-memo","title":"Mark Zuckerberg announces creation of Meta Superintelligence Labs. Read the memo","authors":["CNBC"],"publisher":"CNBC","publicationDate":"2025-06-30","version":"Dated page, retrieved on 2026-10-15 (SHA-256 3c940666760a1f8148ddb1674b0cd4c5f8d28a5eaa32f79fc7052c8cce1ca6b4) through the third-party reader relay r.jina.ai, which renders the publisher’s current page (the publisher refused direct retrieval); not an archive copy.","canonicalUrl":"https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","evidenceUrl":"https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","language":"en","archivedUrl":"https://web.archive.org/web/20261015225529/https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","archivedAt":"2026-10-15","accessedAt":"2026-10-15","reuse":{"attribution":"CNBC, “Mark Zuckerberg announces creation of Meta Superintelligence Labs. Read the memo”, 2025-06-30.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-new-simo-x-2026","title":"Fidji Simo post on X, 15 January 2026 (UTC)","authors":["Fidji Simo"],"publisher":"X","publicationDate":"2026-01-15","version":"Dated post, retrieved on 2026-10-15 (SHA-256 a889f924cb467374d6e5d6865d8f2061e8b77910cf7b67d47902011e23f30890) through X’s own syndication endpoint for the post’s text, which truncates long notes.","canonicalUrl":"https://x.com/fidjissimo/status/2011592010881446116","evidenceUrl":"https://x.com/fidjissimo/status/2011592010881446116","language":"en","archivedUrl":"https://web.archive.org/web/20260115001123/https://twitter.com/fidjissimo/status/2011592010881446116","archivedAt":"2026-01-15","accessedAt":"2026-10-15","reuse":{"attribution":"Fidji Simo, “Fidji Simo post on X, 15 January 2026 (UTC)”, 2026-01-15.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"c2-src-tc260-asgf-v3-pdf","title":"人工智能安全治理框架3.0","authors":["全国网络安全标准化技术委员会"],"publisher":"全国网络安全标准化技术委员会","publicationDate":"2026-09","version":"Dated document, retrieved on 2026-10-27 (SHA-256 cff709d8b1b2a0dfd2ff24df4d6b942b160975331e44c918e14a47d5c86da43d) directly from the publisher. Version: 《人工智能安全治理框架3.0》.pdf.pdf. Bilingual PDF attached to TC260's news item of 14 September 2026 (Chinese text followed by the publisher's English text). Quotes are from the Chinese text. English title (machine translation): “AI Safety Governance Framework 3.0”. Publisher as printed on the inspected expression: 全国网络安全标准化技术委员会 (tc260.org.cn).","canonicalUrl":"https://www.tc260.org.cn/tc260/xwdt1/202609/d513a007d04347f58e483fabaefb34b8/files/%E3%80%8A%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E5%AE%89%E5%85%A8%E6%B2%BB%E7%90%86%E6%A1%86%E6%9E%B63.0%E3%80%8B.pdf.pdf","evidenceUrl":"https://www.tc260.org.cn/tc260/xwdt1/202609/d513a007d04347f58e483fabaefb34b8/files/%E3%80%8A%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E5%AE%89%E5%85%A8%E6%B2%BB%E7%90%86%E6%A1%86%E6%9E%B63.0%E3%80%8B.pdf.pdf","language":"zh-Hans","archivedUrl":"https://web.archive.org/web/20260920144343/https://www.tc260.org.cn/tc260/xwdt1/202609/d513a007d04347f58e483fabaefb34b8/files/%E3%80%8A%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E5%AE%89%E5%85%A8%E6%B2%BB%E7%90%86%E6%A1%86%E6%9E%B63.0%E3%80%8B.pdf.pdf","archivedAt":"2026-09-20","accessedAt":"2026-10-27","reuse":{"attribution":"全国网络安全标准化技术委员会, “人工智能安全治理框架3.0”, 2026-09.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-ch-bis-gp10-guidance-2025","title":"Guidance on Application of General Prohibition 10 (GP10) to People's Republic of China (PRC) Advanced-Computing Integrated Circuits (ICs)","authors":["Bureau of Industry and Security, U.S. Department of Commerce"],"publisher":"Bureau of Industry and Security (bis.gov)","publicationDate":"2025-05-13","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 2770879e8e990d64617a2daed2c924702b0a4562fbab3463537708244c4c03f5). BIS guidance PDF dated May 13, 2025, as served by bis.gov on 2026-10-28 (bis.gov/media/1575 redirects to general-prohibition-10-guidance-may-13-2025.pdf). Law-firm commentary reports that BIS changed the wording of its accompanying press release on May 16, 2025; that release was not retrieved, and whether the PDF itself was revised is not known. Retrieved again at admission on 2026-10-30 (SHA-256 2770879e8e990d64617a2daed2c924702b0a4562fbab3463537708244c4c03f5, the same bytes).","canonicalUrl":"https://www.bis.gov/media/documents/general-prohibition-10-guidance-may-13-2025.pdf","evidenceUrl":"https://www.bis.gov/media/documents/general-prohibition-10-guidance-may-13-2025.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Bureau of Industry and Security, U.S. Department of Commerce, “Guidance on Application of General Prohibition 10 (GP10) to People's Republic of China (PRC) Advanced-Computing Integrated Circuits (ICs)”, 2025-05-13.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-orb-lumen-wp-gh","title":"Why we should train AI in space (White Paper v1.01)","authors":["Ezra Feilden","Adi Oltean","Philip Johnston"],"publisher":"Lumen Orbit, Inc. (now Starcloud), via lumenorbit.github.io","publicationDate":"2024-09","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 90e1bde5b01437d7778e044f8ed95df1c49a63559fc1e35177c94d6443e384f7). PDF v1.01 as served at lumenorbit.github.io/wp.pdf on 2026-10-28; cover line \"White Paper v1.01, September 2024\". Starcloud's current link (starcloud.com/wp) serves v1.03 (starcloudinc.github.io/wp.pdf), which adds a note that Lumen Orbit was renamed Starcloud in 2025; v1.01 is quoted here. The served file cannot authenticate when the v1.01 bytes were first published. Retrieved again at admission on 2026-10-30 (SHA-256 90e1bde5b01437d7778e044f8ed95df1c49a63559fc1e35177c94d6443e384f7, the same bytes). The cross-family check (2026-10-30) opened the Internet Archive capture linked here and found it byte-identical to this retrieval; a capture shows only what the archive held on its date.","canonicalUrl":"https://lumenorbit.github.io/wp.pdf","evidenceUrl":"https://lumenorbit.github.io/wp.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20241224070032/https://lumenorbit.github.io/wp.pdf","archivedAt":"2024-12-24","accessedAt":"2026-10-28","reuse":{"attribution":"Ezra Feilden, Adi Oltean, Philip Johnston, “Why we should train AI in space (White Paper v1.01)”, 2024-09.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-ch-sk-openai-2025","title":"SK Group Partners with OpenAI to Advance Global AI Infrastructure","authors":["SK Group (SK Inc. English newsroom)"],"publisher":"SK Group (SK Inc. English newsroom)","publicationDate":"2025-10-01","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 ce8ca8c2e492c4619f373da358c7191cd3f0dfe3b1b228e832891e2990bdcf66). SK Group English newsroom page as served on 2026-10-28; dateline Seoul, October 1, 2025. SK hynix's own newsroom (news.skhynix.com) was not tried; the SK Group release speaks for SK hynix and SK Telecom. OpenAI's own post on the same announcement could not be retrieved (openai.com refuses direct retrieval from this environment); the cross-family checker should open it or its capture. Retrieved again at admission on 2026-10-30 (SHA-256 8b4cf50a147adca4245785c70d2c0bb6b3200ee2b4932f3b901a6f2c0b07c47c). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://eng.sk.com/news/sk-group-partners-with-openai-to-advance-global-ai-infrastructure","evidenceUrl":"https://eng.sk.com/news/sk-group-partners-with-openai-to-advance-global-ai-infrastructure","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"SK Group (SK Inc. English newsroom), “SK Group Partners with OpenAI to Advance Global AI Infrastructure”, 2025-10-01.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-pw-wnn-ferc-susquehanna-2024","title":"Regulator rejects amended Susquehanna power agreement","authors":["World Nuclear News (World Nuclear Association)"],"publisher":"World Nuclear News (World Nuclear Association)","publicationDate":"2024-11-04","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 44d869b272190ea91a35701bcad5dddd13b8001af54aa3a23b7ca7d6684b1b04). Article as served on 2026-10-28; article:published_time 2024-11-04. The FERC order itself could not be retrieved (www.ferc.gov HTTP 403). Retrieved again at admission on 2026-10-30 (SHA-256 1091501c35b47a4a47fd96f153c71bf48af1168395d2586f08746fa91718a8aa).","canonicalUrl":"https://world-nuclear-news.org/articles/regulator-rejects-amended-susquehanna-power-agreement","evidenceUrl":"https://world-nuclear-news.org/articles/regulator-rejects-amended-susquehanna-power-agreement","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"World Nuclear News (World Nuclear Association), “Regulator rejects amended Susquehanna power agreement”, 2024-11-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-ferc-order-er24-2172","title":"Order Rejecting Amended Interconnection Service Agreement, 189 FERC ¶ 61,078 (Docket ER24-2172)","authors":["Federal Energy Regulatory Commission"],"publisher":"Federal Energy Regulatory Commission","publicationDate":"2024-11-01","version":"Order PDF retrieved at admission on 2026-10-30 (SHA-256 69b04b60cfc5f0150af8cb2f96eabca1c6af9ff74f8bbdbd41a2929b0bf69934, the bytes the cross-family check retrieved).","canonicalUrl":"https://www.ferc.gov/sites/default/files/2024-11/20241101-3061_ER24-2172-000.pdf","evidenceUrl":"https://www.ferc.gov/sites/default/files/2024-11/20241101-3061_ER24-2172-000.pdf","language":"en","accessedAt":"2026-10-30","reuse":{"attribution":"Federal Energy Regulatory Commission, “Order Rejecting Amended Interconnection Service Agreement, 189 FERC ¶ 61,078 (Docket ER24-2172)”, 2024-11-01.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-ab-gov-bwc-msp2025-wp27","title":"Statement by the United Kingdom of Great Britain & Northern Ireland (BWC/MSP/2025/WP.27)","authors":["United Nations (document of the 2025 Meeting of States Parties to the Biological Weapons Convention), submitted by the United Kingdom"],"publisher":"United Nations (document of the 2025 Meeting of States Parties to the Biological Weapons Convention), submitted by the United Kingdom","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 fb628c9a65cfbac99d5250ce48208af6f8ee2ceb4ee867290d2aa600ff83f316). English PDF of document BWC/MSP/2025/WP.27, dated 17 December 2025; 2025 Meeting, Geneva, 15-17 December 2025. Retrieved 2026-10-28 through the documents.un.org symbol access endpoint. Retrieved again at admission on 2026-10-30 (SHA-256 fb628c9a65cfbac99d5250ce48208af6f8ee2ceb4ee867290d2aa600ff83f316, the same bytes).","canonicalUrl":"https://documents.un.org/doc/undoc/gen/g25/211/12/pdf/g2521112.pdf","evidenceUrl":"https://documents.un.org/doc/undoc/gen/g25/211/12/pdf/g2521112.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"United Nations (document of the 2025 Meeting of States Parties to the Biological Weapons Convention), submitted by the United Kingdom, “Statement by the United Kingdom of Great Britain & Northern Ireland (BWC/MSP/2025/WP.27)”.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-ab-hhs-2010-fr","title":"Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA (75 FR 62820)","authors":["Federal Register (Office of the Federal Register, National Archives) via govinfo.gov"],"publisher":"Federal Register (Office of the Federal Register, National Archives) via govinfo.gov","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 d3bfd95a1e3ae4d212505081166f0f8bb7475845e225ae643e6d4a03a93b7e56). Federal Register print of 13 October 2010, Vol. 75, No. 197, pp. 62820-62832, FR Doc. 2010-25728. The notice publishes the final Guidance in full, after the comment responses. Retrieved again at admission on 2026-10-30 (SHA-256 d3bfd95a1e3ae4d212505081166f0f8bb7475845e225ae643e6d4a03a93b7e56, the same bytes).","canonicalUrl":"https://www.govinfo.gov/content/pkg/FR-2010-10-13/pdf/2010-25728.pdf","evidenceUrl":"https://www.govinfo.gov/content/pkg/FR-2010-10-13/pdf/2010-25728.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Federal Register (Office of the Federal Register, National Archives) via govinfo.gov, “Screening Framework Guidance for Providers of Synthetic Double-Stranded DNA (75 FR 62820)”.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-ab-commerce-caisi-bulletin-2025","title":"ICYMI: Statement from U.S. Secretary of Commerce Howard Lutnick on Transforming the U.S. AI Safety Institute into the Pro-Innovation, Pro-Science U.S. Center for AI Standards and Innovation","authors":["United States Department of Commerce"],"publisher":"United States Department of Commerce (GovDelivery bulletin)","publicationDate":"2025-06-04","version":"The bulletin as served, sent 4 June 2025 and carrying a release dated 3 June 2025; retrieved at admission on 2026-10-30. The department’s own release page answered HTTP 403 to the cross-family check. Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://content.govdelivery.com/accounts/USDOC/bulletins/3e381fb","evidenceUrl":"https://content.govdelivery.com/accounts/USDOC/bulletins/3e381fb","language":"en","accessedAt":"2026-10-30","reuse":{"attribution":"United States Department of Commerce, “ICYMI: Statement from U.S. Secretary of Commerce Howard Lutnick on Transforming the U.S. AI Safety Institute into the Pro-Innovation, Pro-Science U.S. Center for AI Standards and Innovation”, 2025-06-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-ab-claude37-card","title":"Claude 3.7 Sonnet System Card","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 82947a46fc2d1b9a1d69501ef973fc3c5b328b703b9a17bd8a143c2672f8a1a6). 42-page PDF as served on 2026-10-28; the request to assets.anthropic.com redirected to https://www-cdn.anthropic.com/9ff93dfa8f445c932415d335c88852ef47f1201e.pdf. The document itself states no publication date; its references are dated February 2025. Retrieved again at admission on 2026-10-30 (SHA-256 82947a46fc2d1b9a1d69501ef973fc3c5b328b703b9a17bd8a143c2672f8a1a6, the same bytes).","canonicalUrl":"https://assets.anthropic.com/m/785e231869ea8b3b/original/claude-3-7-sonnet-system-card.pdf","evidenceUrl":"https://assets.anthropic.com/m/785e231869ea8b3b/original/claude-3-7-sonnet-system-card.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Anthropic, “Claude 3.7 Sonnet System Card”.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}}],"evidence":[{"id":"evidence-guodong-zhang-xai-team-2023","sourceRole":"secondary","support":"supports","locator":"Paragraph below the July 12, 2023 xAI website screenshot","passage":"The reproduced roster includes Guodong Zhang.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-19","note":"Source passage independently inspected by an automated research assistant for this scoped record. Human source review pending."},{"id":"evidence-lineage-kingma-adam","sourceRole":"primary","support":"supports","locator":"v1 PDF p. 1, author block","passage":"The v1 author block credits Diederik P. Kingma. The name-to-entity mapping is separately documented by Kingma’s homepage.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-25","note":"Wording corrected with explicit human authorization on 2026-09-25. Source passage inspected by machine; authorization of the wording does not establish human source inspection."},{"id":"evidence-lineage-kingma-identity","sourceRole":"primary","support":"context","locator":"Page title and heading, Diederik P. (Durk) Kingma / Durk Kingma / Diederik P. Kingma","passage":"Kingma’s own homepage presents Diederik P. Kingma and Durk Kingma as names of the same person.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-25","note":"Supports the identity mapping only; the pinned Adam v1 author block supports authorship. Mutable homepage; no historical expression authenticated. Human source review not asserted."},{"id":"evidence-cn-chatglm-rlhf","sourceRole":"primary","support":"supports","locator":"§1 Introduction, paragraph beginning Following this","passage":"The authors say ChatGPT motivated their subsequent use of SFT and RLHF.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-29","note":"English source paraphrased and checked by machine. Exploratory seed, not editorially accepted. Source reporting is not independent verification of implementation."},{"id":"evidence-supersedes-meta-aasf-v2-0","sourceRole":"primary","support":"supports","locator":"PDF p. 3, Introduction, paragraph beginning “This is the second iteration”; PDF p. 44, Appendix II – Change log","passage":"Version 2 calls itself the second iteration of the Advanced AI Scaling Framework, previously titled the Frontier AI Framework, and its change log lists the February 3, 2025 initial version before the April 7, 2026 entry.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-03","note":"Paraphrase checked by an automated research assistant on 2026-10-03 against the Version 2 PDF, byte-identical to the Internet Archive capture of 2026-04-08. Human editorial review is pending."},{"id":"evidence-supersedes-meta-aasf-v2-0-2","sourceRole":"primary","support":"supports","locator":"Section “Advanced AI Scaling Framework”, first sentence","passage":"Meta’s announcement presents Version 2 as building on its original Frontier AI Framework.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-03","note":"Paraphrase checked by an automated research assistant on 2026-10-03 against the Internet Archive capture of 2026-04-08 and the live page. Human editorial review is pending."},{"id":"evidence-sb-anthropic-fcf-published-1","sourceRole":"primary","support":"supports","locator":"Third paragraph","passage":"Our Frontier Compliance Framework (FCF) is now available to the public","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim excerpt proposed by a research agent and checked by two independent fresh-context verifier agents, both upholding it on 2026-10-04; re-matched mechanically by the integrator against a direct fetch of the cited expression. The page's 'here' link points to trust.anthropic.com (resource 'anthropic-frontier-compliance-framework'). Human editorial review is pending."},{"id":"evidence-sb-anthropic-fcf-published-2","sourceRole":"primary","support":"supports","locator":"Section 'What’s in our Frontier Compliance Framework', final paragraph","passage":"the FCF will serve as our compliance framework for SB 53 and other regulatory requirements","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim excerpt proposed by a research agent and checked by two independent fresh-context verifier agents, both upholding it on 2026-10-04; re-matched mechanically by the integrator against a direct fetch of the cited expression. The sentence begins 'Moving forward,'. Human editorial review is pending."},{"id":"evidence-sb-anthropic-fcf-published-3","sourceRole":"primary","support":"supports","locator":"Section 'What’s in our Frontier Compliance Framework', final paragraph, second sentence","passage":"The RSP will remain our voluntary safety policy","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim excerpt proposed by a research agent and checked by two independent fresh-context verifier agents, both upholding it on 2026-10-04; re-matched mechanically by the integrator against a direct fetch of the cited expression. The sentence continues by saying the RSP reflects what Anthropic believes best practices should be, even where they go beyond or differ from regulatory requirements. Human editorial review is pending."},{"id":"evidence-cs-openai-990-2019-inc-controls-gp-llc-1","sourceRole":"primary","support":"supports","locator":"Schedule R, Part I, line (1)","passage":"Part I lists OpenAI GP LLC (primary activity: research and technology; legal domicile code CA) as a disregarded entity whose direct controlling entity is OpenAI Inc.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-04","note":"Editorial paraphrase, re-checked on 2026-10-04 by the integrating session against its own direct fetch of the cited expression. The integrating session wrote or selected this passage from its own reading of the source; it does not repeat a passage from the week 4b research pass. Human editorial review is pending."},{"id":"evidence-prov-fasc-2024-intolerable-thresholds-1","sourceRole":"primary","support":"supports","locator":"Outcome 1, commitment II, first sentence","passage":"at which severe risks posed by a model or system, unless adequately mitigated, would be deemed intolerable","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim quote, re-checked on 2026-10-04 by the integrating session against its own direct fetch of the cited expression (mechanical match after whitespace normalisation). Also matched in the Internet Archive capture of 2024-05-21. The passage was first proposed and machine-verified in the week 4b research pass. Human editorial review is pending."},{"id":"evidence-ts-a2-turing-1950-context","sourceRole":"primary","support":"supports","locator":"Section 7, Learning Machines, printed p. 456 (scan 24), opening paragraphs","passage":"Instead of programming an adult mind directly, Turing proposes programming a child-like starting state and educating it, comparing child programmes by experiment.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-06","note":"Paraphrase from the pre-1996 candidate search’s inspection of the page image; not re-read for this collection."},{"id":"evidence-ts-a2-turing-1950-context-2","sourceRole":"primary","support":"supports","locator":"Section 6 (5), printed p. 449 (scan 17), paragraph on a machine being its own subject matter","passage":"It may be used to help in making up its own programmes, or to predict the effect of alterations in its own structure. By observing the results of its own behaviour it can modify its own programmes so as to achieve some purpose more effectively.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-06","note":"Read on the page image of printed p. 449 (scan 17) for this collection."},{"id":"evidence-yk-t14-s3","sourceRole":"primary","support":"supports","locator":"message body, paragraph beginning 'CFAI did *not* propose to explicitly design a Sysop scenario'","passage":"CFAI … *did* say … build right away, and work out the FAI part as we go along","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-07","note":"Candidate pack passage T14-s3. The quotation (18 words; trimmed 2026-10-13 from 24 words) matched the retrieved expression (SHA-256 08cb8b0fc83f702f32882f8b3eba0724135658d8406a8c761c4617d8601412fb) when checked on 2026-10-07.","contextOnly":{"reason":"An earlier view that only partly matches the take; never show alone","requires":"The following paragraph of the same message, described in the record's statement, in which he says he realised in 2002 that almost every AI design would kill him."}},{"id":"evidence-omni-art56-adequacy-assessment-1","sourceRole":"primary","support":"supports","locator":"Article 1, point (21)","passage":"The Commission, taking utmost account of the opinion of the Board, shall assess whether the codes of practice cover the obligations provided for in Articles 53 and 55","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-10","note":"Verbatim quote, matched mechanically (after whitespace normalisation) by the proposing session against the text retrieved on 2026-10-10. Checked by fresh-context machine audit on 2026-10-10 (review/frontier-governance-audit-2026-10-10.md). Human editorial review is pending."},{"id":"evidence-omni-art56-adequacy-assessment-2","sourceRole":"primary","support":"supports","locator":"Article 1, point (21)","passage":"The Commission shall publish its assessment of the adequacy of the codes of practice.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-10","note":"Verbatim quote, matched mechanically (after whitespace normalisation) by the proposing session against the text retrieved on 2026-10-10. Checked by fresh-context machine audit on 2026-10-10 (review/frontier-governance-audit-2026-10-10.md). Human editorial review is pending."},{"id":"evidence-omni-art56-adequacy-assessment-3","sourceRole":"primary","support":"context","locator":"Recital (41)","passage":"the codes of practice referred to in Article 50(7) and Article 56(6) have limited legal effect, and in particular do not grant a presumption of conformity","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-10","note":"Verbatim quote, matched mechanically (after whitespace normalisation) by the proposing session against the text retrieved on 2026-10-10. Recital text, quoted as the Regulation’s own description of codes of practice. Checked by fresh-context machine audit on 2026-10-10 (review/frontier-governance-audit-2026-10-10.md). Human editorial review is pending."},{"id":"evidence-ec-m09-04","sourceRole":"primary","support":"supports","locator":"Part II, sec. 8 'Has there been an increase in the rate of technological change?', printed p. II-105","passage":"It is easy to find statements by economists and others asserting that the rate of technological change in the postwar period is much more rapid than before the war","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M09-04 (research/economics-mainstream). The quotation (29 words) matched the retrieved current copy (SHA-256 85b1fb59b997e8cbc0f7255076a55f68008955a2b10523b375ac0ae58628572c) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day.","contextOnly":{"reason":"Fresh-context audit (A2) judged that the quotation misleads when shown alone.","requires":"Truncated before 'and it is equally possible to find statements asserting the opposite'. Mansfield then examines the evidence. Shown alone it reads as a consensus claim."}},{"id":"evidence-ec-m05-00","sourceRole":"primary","support":"supports","locator":"standfirst on the publisher page","passage":"Can industry afford to buy them, and will they cause widespread technological unemployment?","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M05-00 (research/economics-mainstream). The quotation (13 words) matched the retrieved current copy (SHA-256 f9161d3789c10d798b9d27dad8762bef7874d09449ea5dad786db5ae180ad24f) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day, whose SHA-256 differs from the recorded one.","contextOnly":{"reason":"Fresh-context audit (A2) judged that the quotation misleads when shown alone.","requires":"Magazine standfirsts are normally written by editors; attribution to Leontief is unverified. It frames the article's questions and gives none of its answers."}},{"id":"evidence-ec-m07-09","sourceRole":"primary","support":"supports","locator":"ETLA publication page, abstract","passage":"The estimated impacts do not necessarily imply future mass unemployment, since the approach employed does not take into account changes in the task content within occupations or the evolution in the mix of occupations.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M07-09 (research/economics-mainstream). The quotation (34 words) matched the retrieved current copy (SHA-256 f14a8103d4eb6603c6092abe47665ed38e408020064249fedf4c0bf5747d7f9e) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day, whose SHA-256 differs from the recorded one."},{"id":"evidence-ec-e03-05","sourceRole":"primary","support":"supports","locator":"NBER WP 24174, footnote 31, p. 35","passage":"However, if the gap between humans and super‐intelligent AI entities grows too large, it may be impossible for humans to continue to exert control, just like a two‐year old would not be able to effectively exert property rights over adults.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-16","note":"Candidate pack passage E03-05 (research/economics-mainstream/esoteric). The quotation (40 words) matched the retrieved current copy (SHA-256 aeddf962b5c488cae68b7d48cef964de22379c78e75579c5793f0df71d5e9f40) in the pack's verification of 2026-10-15. Re-checked at admission on 2026-10-16 against a copy retrieved that day."},{"id":"evidence-ec-e03-06","sourceRole":"primary","support":"supports","locator":"NBER WP 24174, \"Second Scenario: Artificially Intelligent Agents and the Return of Malthus\", p. 36","passage":"Taken together, these two dynamics imply – unsurprisingly – that humans will lose the Malthusian race in the long run, unless counteracting steps are taken","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-16","note":"Candidate pack passage E03-06 (research/economics-mainstream/esoteric). The quotation (25 words) matched the retrieved current copy (SHA-256 aeddf962b5c488cae68b7d48cef964de22379c78e75579c5793f0df71d5e9f40) in the pack's verification of 2026-10-15. Re-checked at admission on 2026-10-16 against a copy retrieved that day.","contextOnly":{"reason":"A model scenario, not the author's own forecast (maintainer decision D3).","requires":"The model's conditions and the author's own assessment."}},{"id":"evidence-ec-e04-14","sourceRole":"primary","support":"supports","locator":"pdf p.30; §6 Conclusion, 'Preparing for dangerous AI'","passage":"Policy responses center on established approaches: advancing technical AI safety research to ensure system alignment and robustness, and developing governance frameworks capable of managing rapid capability transitions while maintaining incentives for innovation.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-16","note":"Candidate pack passage E04-14 (research/economics-mainstream/esoteric). The quotation (32 words) matched the retrieved current copy (SHA-256 06dda75180302720e0843a180c00d04fa99e38db0f5d6d4e06d7f29388a115bb) in the pack's verification of 2026-10-15. Re-checked at admission on 2026-10-16 against a copy retrieved that day."},{"id":"lg-ev-goo-vaswani-essential-ai-ceo-2023","sourceRole":"primary","support":"supports","locator":"Paragraph 5","passage":"said Vaswani, CEO of Essential AI","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (6 words) matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read through the third-party reader relay r.jina.ai, which renders the publisher’s current page (the publisher refused direct retrieval); not an archive copy. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"reader_relay"},{"id":"lg-ev-mx-sun-msl-2025","sourceRole":"secondary","support":"supports","locator":"Memo, list of new team members","passage":"The memo lists Pei Sun as working on post-training, coding and reasoning for Gemini at Google DeepMind, previously on Waymo perception models.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Pei Sun” matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read through the third-party reader relay r.jina.ai, which renders the publisher’s current page (the publisher refused direct retrieval); not an archive copy. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The paraphrase is Selekoa’s; only the anchor is the source’s wording. A cross-family machine check (2026-10-16/17) re-read the record’s sources and agreed with the record. No human editorial review.","retrievalRoute":"reader_relay"},{"id":"lg-ev-new-zoph-openai-return-2026","sourceRole":"primary","support":"supports","locator":"X post","passage":"Simo welcomed Barret Zoph, Luke Metz and Sam Schoenholz back to OpenAI and wrote that Zoph would report to her.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “welcome Barret Zoph, Luke Metz” matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read through X’s own syndication endpoint for the post’s text, which truncates long notes. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The paraphrase is Selekoa’s; only the anchor is the source’s wording. No human editorial review.","retrievalRoute":"x_syndication"},{"id":"ev-cmp-fsf-v3-0-d6-discretion-1","sourceRole":"primary","support":"supports","locator":"PDF p. 7 (printed p. 7), §1.6 Summary of Risk Acceptance Criteria, misuse-CCL bullet (begins PDF p. 6), second sub-bullet, from “or we otherwise assess”","passage":"or we otherwise assess that the level of security applied is adequate, e.g. if they match or exceed the level of security applied to other models with similar capabilities or risk profiles, or we assess that the benefits of the open release of model weights outweigh the risks.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Existing dataset passage for this edition, which contains the comparison’s shorter quotation, first checked on 2026-10-03. The comparison’s quotation was checked by an automated research assistant on 2026-10-27 against the publisher copy, whose SHA-256 matches the copy retrieved on 2026-10-15. Alternative to reaching the recommended security level; examples given include assessing that the benefits of open release of weights outweigh the risks. Human editorial review is pending."},{"id":"evidence-ap-daron-acemoglu-acemoglu-restrepo-2016-race-between-machine-and-man","sourceRole":"primary","support":"supports","locator":"PDF text layer, first two pages: author name on the title page or first page","passage":"Daron Acemoglu","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Credited name checked by an automated research assistant on 2026-10-27 in a direct retrieval and checked by a reviewer from a different model family on 2026-10-27. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"evidence-ap-robin-hanson-extropians-1998-09-24-hanson","sourceRole":"primary","support":"supports","locator":"Message header, sender name (address not reproduced)","passage":"Credited as “Robin Hanson” (message header, sender name (address not reproduced)).","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-27","note":"Credited name checked by an automated research assistant on 2026-10-27 in a direct retrieval and checked by a reviewer from a different model family on 2026-10-27. Given as a labelled paraphrase because a quotation would exceed the short-quotation limit for this document. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"evidence-ap-wassily-leontief-leontief-duchin-1984-impacts-of-automation","sourceRole":"primary","support":"supports","locator":"PDF page 2, report title page, “by” line","passage":"Wassily Leontief and Faye Duchin","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Passage and locator as inspected by a reviewer from a different model family on 2026-10-27 in a direct retrieval of the Internet Archive microfiche scan; the title page was checked visually because the OCR text misreads the names; the preparer’s matcher had found the name only outside the byline. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"ev-cmp-rsp-v3-0-d1-persuasion","sourceRole":"primary","support":"supports","locator":"RSP · v3.0: Whole text layer, PDF pp. 1–19 (all sections, appendices and changelog read).","passage":"Within the inspected scope, no passage meeting the comparison question “Persuasion or manipulation” was found.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-26","note":"Absence recorded by an automated research assistant on 2026-10-15 for the framework comparison, audited in a fresh context, checked by a reviewer from a different model family on 2026-10-15 and re-checked on 2026-10-26. Human editorial review is pending."},{"id":"ev-cmp-rsp-v3-3-d1-autonomy-1","sourceRole":"primary","support":"supports","locator":"PDF p. 8, §1, threshold table, row 'Automated R&D in key domains', left column, first sentence","passage":"The “Automated R&D in key domains” row of the threshold table concerns AI systems capable of fully automating, or dramatically speeding up, the work of large teams of leading human researchers in domains where rapid progress could threaten international security or quickly disrupt the global balance of power.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-28","note":"Explicit paraphrase by an automated research assistant, corrected on 2026-10-28 after a blind check of all 287 comparison paraphrases by a reviewer from a different model family on 2026-10-28 (research/frontier-framework-comparison/codex-paraphrases-2026-10-28.md), which found that it left out the restriction to domains where fast progress could threaten international security or the global balance of power. The correction was written from the passage as re-read in the publisher copy retrieved on 2026-10-28, whose SHA-256 matches the copy retrieved on 2026-10-15. Given as a paraphrase to keep the edition within the quotation limit. The examples given are energy, robotics, weapons development and AI. Human editorial review is pending."},{"id":"c2-ev-governance-tc260-asgf-v3-agent-risk-framework-1","sourceRole":"primary","support":"supports","locator":"§4.4.1","passage":"建立智能体风险管理框架，系统分析智能体的安全风险，研究提出安全风险防范措施和应对建议","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (40 words, each CJK character counted as one) matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. English machine translation, not verified by a person: “establish a risk management framework for intelligent agents, systematically analyse the safety risks of agents, and research and propose safety risk prevention measures and response recommendations”. No human editorial review. On 2026-10-30, by the maintainer’s decision of that day, the quotation was corrected to the archive capture of 2026-09-20 (full-width commas as in the capture), which a cross-family archive check read on 2026-10-28; the proposition is unchanged. The retrieval route is that of the earlier check.","retrievalRoute":"direct"},{"id":"ev-cmp-meta-faf-v1-1-d1-autonomy-1","sourceRole":"primary","support":"supports","locator":"PDF p. 14 (printed p. 14), §3.4.1, cyber outcomes table, enabling capabilities column","passage":"In its cyber outcomes table, the edition lists autonomous cyber operations among the enabling capabilities.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-27","note":"Explicit paraphrase, written on 2026-10-28 by an automated research assistant from the passage that a reviewer from a different model family quoted and located on 2026-10-27 in the Internet Archive capture of 2025-02-13, which it retrieved and hashed (decoded PDF SHA-256 9fa301df…cc21; research/frontier-framework-comparison/codex-meta-v1-1-2026-10-27.md). The edition could not be retrieved again for the paraphrase. Human editorial review is pending."},{"id":"evidence-fl-link-rsp-v3-3-sb53-1","sourceRole":"primary","support":"supports","locator":"Footnote 1 (Introduction), PDF p. 4","passage":"Where laws such as California SB 53 define this or similar terms with specific thresholds, we address those requirements in separate compliance frameworks.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (23 words) matched the text retrieved for the candidate pack on 2026-10-27 and again at admission on 2026-10-28, in a fresh retrieval whose SHA-256 equals the pack’s. Read directly from cdn.sanity.io. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. Same sentence as in v3.0. A cross-family machine check (2026-10-28, item 117) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-ch-bis-gp10-guidance-ascend-list-1","sourceRole":"primary","support":"supports","locator":"p. 1, list","passage":"PRC 3A090 ICs Subject to Presumption of GP10 Restrictions Huawei Ascend 910B Huawei Ascend 910C Huawei Ascent 910D","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (18 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-ch-bis-gp10-guidance-ascend-list-2","sourceRole":"primary","support":"supports","locator":"p. 1, illustrative list paragraph","passage":"This list is not exhaustive","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (5 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc7-lumen-2024-1","sourceRole":"primary","support":"supports","locator":"PDF p. 10, \"Orbit\"","passage":"shielding is required to reduce radiation-induced effects, including latch-up, transients, and total ionizing dose (TID) effects in sensitive components such as storage and power delivery","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (25 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 97) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc7-lumen-2024-2","sourceRole":"primary","support":"supports","locator":"PDF p. 11, \"Orbit\"","passage":"Therefore the mass of shielding needed per compute unit decreases linearly with container size.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (14 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 97) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-1","sourceRole":"secondary","support":"context","locator":"paragraph beginning “The US Federal Energy Regulatory Commission…”","passage":"rejecting the amended Susquehanna Interconnection Service Agreement","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (7 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 1091501c35b47a4a47fd96f153c71bf48af1168395d2586f08746fa91718a8aa; the bytes differ from the pack’s retrieval). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 44) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-2","sourceRole":"secondary","support":"context","locator":"paragraph beginning “The US Federal Energy Regulatory Commission…”","passage":"the agreement rejected is described as between PJM Interconnection, PPL Electric Utilities and Talen Energy Corporation.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The paraphrase is Selekoa’s, written from the text retrieved for the candidate pack on 2026-10-28; it is not a quotation and was not machine-matched. The cross-family check found it faithful. A cross-family machine check (2026-10-29/30, item 44) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-3","sourceRole":"secondary","support":"context","locator":"paragraph beginning “PJM is the regional transmission organisation…”","passage":"to increase from 300 MW to 480 MW the amount of load it is allowed to transfer from Susquehanna","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (19 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 1091501c35b47a4a47fd96f153c71bf48af1168395d2586f08746fa91718a8aa; the bytes differ from the pack’s retrieval). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 44) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-4","sourceRole":"secondary","support":"context","locator":"paragraph beginning “FERC commissioners have now voted 2-1…”","passage":"FERC commissioners have now voted 2-1 against the proposal","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (9 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 1091501c35b47a4a47fd96f153c71bf48af1168395d2586f08746fa91718a8aa; the bytes differ from the pack’s retrieval). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 44) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-5","sourceRole":"secondary","support":"context","locator":"Talen statement quoted in the report","passage":"the report quotes Talen as saying that the existing ISA allows for 300 MW of co-located load at Susquehanna.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The paraphrase is Selekoa’s, written from the text retrieved for the candidate pack on 2026-10-28; it is not a quotation and was not machine-matched. The cross-family check found it faithful. A cross-family machine check (2026-10-29/30, item 44) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-6","sourceRole":"primary","support":"supports","locator":"189 FERC ¶ 61,078, PDF p. 1, paragraph 1","passage":"to increase the amount of Co-Located Load from 300 megawatts (MW) to 480 MW","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Added at the admission of the compute, power and orbit pack (research/compute-power-orbit). The regulator’s own order (cross-family check, item 44). The quotation (14 words) matched a direct retrieval of 2026-10-30 (SHA-256 69b04b60cfc5f0150af8cb2f96eabca1c6af9ff74f8bbdbd41a2929b0bf69934). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-ferc-susquehanna-isa-rejected-7","sourceRole":"primary","support":"supports","locator":"189 FERC ¶ 61,078, PDF p. 1, paragraph 1","passage":"In this order, we reject the Amended ISA.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Added at the admission of the compute, power and orbit pack (research/compute-power-orbit). The decision as the order states it. The quotation (8 words) matched a direct retrieval of 2026-10-30 (SHA-256 69b04b60cfc5f0150af8cb2f96eabca1c6af9ff74f8bbdbd41a2929b0bf69934). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-ch-skhynix-openai-loi-1","sourceRole":"primary","support":"supports","locator":"body, paragraph 2","passage":"SK hynix signed a Letter of Intent (LOI) with OpenAI to supply High Bandwidth Memory (HBM), for the construction of global AI data centers under the Stargate project.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (28 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 8b4cf50a147adca4245785c70d2c0bb6b3200ee2b4932f3b901a6f2c0b07c47c; the bytes differ from the pack’s retrieval). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 68) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-ch-skhynix-openai-loi-2","sourceRole":"primary","support":"supports","locator":"body, paragraph 2","passage":"SK hynix will establish a production system capable of promptly meeting OpenAI’s HBM demand for up to 900,000 DRAM wafers per month. It is more than double the current HBM industry capacity","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (32 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 8b4cf50a147adca4245785c70d2c0bb6b3200ee2b4932f3b901a6f2c0b07c47c; the bytes differ from the pack’s retrieval). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 68) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-scr-hhs-2010-effective-date-1","sourceRole":"primary","support":"supports","locator":"DATES, 75 FR 62820","passage":"The Guidance is effective on October 13, 2010.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (8 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-30, item 55) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-gov-bwc-uk-wp27-biotech-ai-threats-1","sourceRole":"primary","support":"supports","locator":"Paragraph 5","passage":"At present, technology is outstripping diplomacy. Novel and hybrid threats to our collective security have emerged, including using biotechnology and AI as new kinds of weapons of war.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (28 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. The paragraph continues that these threats come in many different forms and from many different quarters. A cross-family machine check (2026-10-30, item 90) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-gov-us-commerce-caisi-bulletin-1","sourceRole":"primary","support":"supports","locator":"Bulletin body, opening paragraph of the release of 3 June 2025","passage":"announced his plans to reform the agency formerly known as the U.S. AI Safety Institute into the Center for AI Standards and Innovation (CAISI)","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Added at the admission of the AI and biosecurity pack (research/ai-bio). The documentary statement the cross-family check located for the June 2025 name, after the department’s own release page refused retrieval. The quotation (24 words) matched a direct retrieval of 2026-10-30 (SHA-256 86242f337627e2842631ea8e564cc742e7e90118527ee3e0aa567ed71b232222). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-claude37-asl2-determination-1","sourceRole":"primary","support":"supports","locator":"PDF p. 23, §7 RSP Evaluations, heading line","passage":"RSP safeguards required for Claude 3.7 Sonnet: ASL-2","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (8 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-30, item 47) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-claude37-asl2-determination-2","sourceRole":"primary","support":"supports","locator":"PDF p. 24, §7.1 CBRN Evaluations, conclusion paragraph","passage":"the results from our evaluations suggest improved performance in all domains, including some uplift in CBRN evaluations","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (17 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-30, item 47) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-claude37-asl2-determination-3","sourceRole":"primary","support":"supports","locator":"PDF p. 24, §7.1, conclusion paragraph","passage":"based on what we observed in our recent CBRN testing, we believe there is a substantial probability that our next model may require ASL-3 safeguards","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (25 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-30, item 47) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-claude37-asl2-determination-4","sourceRole":"primary","support":"context","locator":"PDF p. 23, §7.1, 'Threshold and evaluations' paragraph","passage":"Under our RSP framework, we consider some of the specific evaluation thresholds and results in the CBRN domain sensitive, and therefore we do not share them publicly.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (27 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. The card's stated limit on what it publishes. A cross-family machine check (2026-10-30, item 47) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"}],"at":{"entities":[0,1,3,45,48,51,70,78,89,98,141,194,199,200,205,206,207,209,252,258,306,337,363,483,492,517,520,522,538,540,573,579,580,583,587,588,601,636,669,741,765,848,884,894,1004,1088,1091,1121,1166,1195,1203,1215,1221,1236],"claims":[182,194,386,393,625,629,721,771,1082,1212,1451,1473,1477,1522,1541,1671,1781,1838,2113,2170,2250,2284,2532,2567,2844,2850,2926,3069,3096,3191,3216,3283,3350,3362,3404],"sources":[22,39,42,46,75,89,173,178,180,182,203,207,239,258,309,450,458,490,492,526,532,533,536,550,634,700,730,834,912,920,951,1012,1031,1042,1044,1061,1081],"evidence":[201,213,245,409,417,418,741,742,743,752,904,978,979,1318,1488,1489,1490,1808,1839,1843,1896,1897,1926,2106,2289,2364,2740,2815,2897,2931,3182,3223,3642,3653,3780,4025,4026,4071,4072,4309,4310,4311,4312,4313,4314,4315,4382,4383,4500,4599,4613,4683,4684,4685,4686]},"version":"1lg47t7"}