{"entities":[{"id":"openai","name":"OpenAI"},{"id":"anthropic","name":"Anthropic"},{"id":"amazon","name":"Amazon"},{"id":"rsp-2023","name":"Responsible Scaling Policy (2023)"},{"id":"tom-brown","name":"Tom Brown"},{"id":"rsp-v3-0","name":"Anthropic Responsible Scaling Policy — v3.0"},{"id":"jan-leike","name":"Jan Leike"},{"id":"human-preferences-paper","name":"Deep reinforcement learning from human preferences"},{"id":"self-critique-paper","name":"Self-critiquing models for assisting human evaluators"},{"id":"weak-strong-paper","name":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision"},{"id":"gym-announcement","name":"OpenAI Gym Beta (2016)"},{"id":"meta","name":"Meta"},{"id":"zhipu","name":"Zhipu AI / 智谱"},{"id":"wang-shaolan","name":"Wang Shaolan / 王绍兰"},{"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":"openai-lp","name":"OpenAI LP"},{"id":"openai-gp-llc","name":"OpenAI GP LLC"},{"id":"anthropic-coi-2023-09-21","name":"Amended and Restated Certificate of Incorporation of Anthropic, PBC (filed 2023-09-21)"},{"id":"anthropic-coi-2024-01-29","name":"Amended and Restated Certificate of Incorporation of Anthropic, PBC (filed 2024-01-29)"},{"id":"wiener-1960-some-moral-technical","name":"Some Moral and Technical Consequences of Automation"},{"id":"moravec-1989-genetic-takeover","name":"Human Culture—A Genetic Takeover Underway"},{"id":"fast-or-hard-takeoff","name":"Fast or hard takeoff"},{"id":"cn-generative-ai-interim-measures-2023","name":"Interim Measures for the Management of Generative Artificial Intelligence Services (生成式人工智能服务管理暂行办法), Order No. 15"},{"id":"long-run-history-suggests-acceleration","name":"Long-run history suggests a further acceleration"},{"id":"clancy-besiroglu-2023-great-inflection","name":"The Great Inflection? A Debate About AI and Explosive Growth"},{"id":"benzell-et-al-2015-robots-are-us","name":"Robots Are Us: Some Economics of Human Replacement"},{"id":"brynjolfsson-rock-syverson-2017-productivity-paradox","name":"Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics"},{"id":"baidu","name":"Baidu / 百度"},{"id":"h-company","name":"H Company"},{"id":"jia-lei","name":"Jia Lei"},{"id":"julien-perolat","name":"Julien Perolat"},{"id":"karen-simonyan","name":"Karén Simonyan"},{"id":"lilian-weng","name":"Lilian Weng"},{"id":"microsoft-ai","name":"Microsoft AI"},{"id":"sensetime","name":"SenseTime / 商汤科技"},{"id":"yun-yeyi","name":"Yun Yeyi"},{"id":"seth-g-benzell","name":"Seth G. Benzell"},{"id":"daniel-rock","name":"Daniel Rock"},{"id":"cn-ai-content-labelling-measures-2025","name":"Measures for Labelling Artificial Intelligence-Generated Synthetic Content (人工智能生成合成内容标识办法)"},{"id":"ascend-results-release-2024","name":"Thales Alenia Space reveals results of ASCEND feasibility study on space data centers"},{"id":"oc2-radiative-cooling-advantage","name":"Radiative cooling in orbit as an advantage"},{"id":"vct-benchmark-paper-2025","name":"Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark (arXiv:2504.16137)"}],"claims":[{"id":"amazon-anthropic-new-2024","statement":"Anthropic announced a new $4 billion Amazon investment, which would bring the total to $8 billion.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Future cumulative total is not encoded as this transaction’s amount or proof of same-day closing.","evidenceIds":["evidence-amazon-anthropic-new-2024"],"attributes":{"transactionType":"investment","relationshipContext":"external","transactionStage":"announced","amount":{"value":4000000000,"currency":"USD","bound":"exact","scope":"transaction"}}},{"id":"lineage-jan-leike-preferences","statement":"Deep reinforcement learning from human preferences names Jan Leike 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-jan-leike-preferences"]},{"id":"lineage-ii-weak-cites-critique","statement":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision cites Self-critiquing models for assisting human evaluators.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Work-level citation; the cited revision is unspecified. Does not establish agreement, causal influence, implementation or institutional adoption. The paragraph contrasts improving supervision with generalizing beyond it; no method-use edge is justified.","evidenceIds":["evidence-lineage-ii-weak-cites-critique"]},{"id":"atlas-gym","statement":"OpenAI published the Gym beta announcement, introducing a toolkit for comparing reinforcement-learning algorithms.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Publisher account of its own work; publication does not independently establish implementation, effectiveness or priority.","evidenceIds":["evidence-atlas-gym"]},{"id":"issued-meta-faf-v1-1","statement":"Meta published its Frontier AI Framework, Version 1.1, identified by its announcement date of 2025-02-03.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Publication means a public document expression exists. Publication is not evidence that its requirements were implemented or effective in practice. A revised expression under the same Version 1.1 label followed by 2025-03-28 without a located announcement.","evidenceIds":["evidence-issued-meta-faf-v1-1","evidence-issued-meta-faf-v1-1-2"]},{"id":"prov-meta-aasf-v2-0-high-deployment-measure","statement":"Advanced AI Scaling Framework v2 states that the measure at the high risk threshold is to proceed with deployment of the Frontier AI only if sufficient mitigations are defined, implemented and validated to reduce risk to that of a moderate or lower model.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the edition states, quoted from the cited expression. It is not evidence of implementation, compliance or effectiveness.","evidenceIds":["evidence-prov-meta-aasf-v2-0-high-deployment-measure"],"attributes":{"provision":{"kind":"required_action","key":"meta-high-deployment-measure"}}},{"id":"cs-openai-990-2019-gp-llc-controls-lp","statement":"OpenAI, Inc.’s Form 990 for tax year 2019 lists OpenAI LP (EIN 83-1960637) as a related organisation taxable as a partnership whose direct controlling entity is OpenAI GP LLC, and answers “Yes” to whether the filing organisation is a general or managing partner of it.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A self-reported tax schedule; the form’s terms are not interpreted under partnership law. OpenAI GP LLC is reported on the same schedule as a disregarded entity of OpenAI Inc; the two records are not joined into a chain of control. The partnership agreement was not inspected.","evidenceIds":["evidence-cs-openai-990-2019-gp-llc-controls-lp-1"],"attributes":{"control":{"right":"controls","instrument":"Form 990, tax year 2019, Schedule R Part III: direct controlling entity of a related partnership","stage":"effective"}}},{"id":"provchange-anthropic-coi-2024-01-29-class-t-phase-in","statement":"Compared with the certificate filed on 21 September 2023, the certificate filed on 29 January 2024 moves the funding-based Class T phase-in triggers from the board’s determination itself to four months (second phase) and eight months (final phase) after the board determines that $5 billion or $6 billion has been raised. The outside dates of 24 May 2026 and 24 May 2027 are unchanged.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: Article VIII(B)6(iv) and (vi) of the certificate filed on 2023-09-21 tie the Final and Second Phase-In Dates to “such time as the Board of Directors determines” the funding thresholds. The same definitions in the certificate filed on 2024-01-29 read “eight months following” and “four months following the date on which the Board of Directors determines”. Reasoning: a direct comparison of the two definitions as read from the page images; neither filing contains a change log.","editorialState":"seed","qualification":"A textual comparison of two filed versions. It does not establish when any phase-in date occurred, which depends on board determinations that were not published, and no certificate filed between these two was inspected.","evidenceIds":["evidence-provchange-anthropic-coi-2024-01-29-class-t-phase-in-1","evidence-provchange-anthropic-coi-2024-01-29-class-t-phase-in-2"],"attributes":{"changeBasis":"document_comparison","provision":{"kind":"threshold","key":"class-t-phase-in"}}},{"id":"ts-a4-wiener-1960-context","statement":"Wiener’s 1960 article warns that machines may act faster than people can gather the information needed to stop them.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Not a takeoff timescale.","evidenceIds":["evidence-ts-a4-wiener-1960-context"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"The speed discussed is of machine operation and decision, not a rapid transition in intelligence.","claimed":{"date":"1959-12-27","note":"Printed p. 1355 says the article is adapted from a lecture at the AAAS meeting in Chicago on this day. The lecture text was not retrieved, so the wording is not dated to it."}}}},{"id":"ts-a4-moravec-1989-context","statement":"Moravec’s chapter says machines are visibly overtaking people.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"No timescale for a transition.","evidenceIds":["evidence-ts-a4-moravec-1989-context"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"No transition period of hours to months.","claimed":{"date":"1988-07","note":"The author’s file is dated July 1988."},"attested":{"date":"2001-03-06","note":"Author-hosted file; the earliest capture, of 2001-03-06, is byte-identical to the copy retrieved in 2026. Not compared with the printed chapter."}}}},{"id":"ec-g03-clancy-besiroglu-2023-great-inflection-a","statement":"Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G03. Speaker: Tamay Besiroglu. Points to faster growth after the Neolithic and Industrial Revolutions, and to double-digit catch-up growth, as precedent.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The quotation is a turn by Tamay Besiroglu in a dialogue.","evidenceIds":["evidence-ec-g03-f"],"attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","claimed":{"date":"2023","note":"\"back-and-forths ... in spring of 2023\"; the original exchange was not retrieved and the published dialogue may be edited."},"attested":{"date":"2023-12-23","note":"Only retrieved capture."},"tier":{"value":2,"basis":"Tier 2 (unchanged). Magazine debate (Asterisk). The magazine introduces Matt Clancy as an economist and Tamay Besiroglu as a research scientist. Judgement recorded: no turn is itself a substantive scholarly argument; each answers the previous turn and argues by verbal sketches of others' models, with no model or derivation of its own, so the debate rule applies to both speakers."}}}},{"id":"lg-cn-yun-yeyi-sensetime-2017-2022","statement":"MiniMax’s Hong Kong prospectus of 31 December 2025 states that Yun Yeyi held positions at SenseTime from September 2017 to January 2022, latterly as director of innovative business.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A biography in MiniMax’s filing, not a SenseTime record. The interval joins the listed positions; no reason for leaving is stated.","evidenceIds":["lg-ev-cn-yun-yeyi-sensetime-2017-2022"],"attributes":{"role":"various positions; director of innovative business (2021-01 to 2022-01)"}},{"id":"lg-deal-simonyan-microsoft-ai-chief-scientist-2024","statement":"Microsoft announced on 19 March 2024 that Karén Simonyan was joining Microsoft AI as Chief Scientist, reporting to Mustafa Suleyman.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An announced appointment; the record does not establish a start date or a continuing role.","evidenceIds":["lg-ev-deal-simonyan-microsoft-ai-chief-scientist-2024"],"attributes":{"activityStage":"announced","role":"Chief Scientist, Microsoft AI"}},{"id":"lg-gw-perolat-h-company-departure-2024","statement":"Sifted reported on 24 August 2024 that Julien Perolat, a co-founder of H Company, was leaving the company.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting (Sifted); most of the article was behind a paywall, so only its headline, standfirst and opening paragraph were checked. The company statement it relays (a LinkedIn post) was not retrieved, and the reason the report gives is not recorded. It does not establish a last working day.","evidenceIds":["lg-ev-gw-perolat-h-company-departure-2024-a","lg-ev-gw-perolat-h-company-departure-2024-b"],"attributes":{"role":"co-founder"}},{"id":"lg-oai-weng-departure-2024","statement":"Lilian Weng posted on X on 8 November 2024 that she had decided to leave OpenAI after almost seven years.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A first-person announcement; the note she shared with her team was posted as an image and is not used. Her later movements are not recorded here.","evidenceIds":["lg-ev-oai-weng-departure-2024","lg-ev-oai-weng-departure-2024-title"],"attributes":{"role":"Not stated in the post (title Head of Safety Systems from “OpenAI Board Forms Safety and Security Committee” (OpenAI), 28 May 2024)"}},{"id":"lg-gc-jia-lei-baidu-application-model-2025","statement":"The Paper reported on 25 November 2025 that its reporter had learned that day of a Baidu announcement creating a new Application Model R&D Department headed by Jia Lei.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting of a Baidu announcement that was not retrieved. A continuing role is not established. English rendering of the Chinese passage is a machine translation; no human translation review.","evidenceIds":["lg-ev-gc-jia-lei-baidu-application-model-2025"],"attributes":{"role":"负责人 (head), 应用模型研发部 (Application Model R&D Department)"}},{"id":"lg-ant-brown-chief-compute-2026","statement":"Anthropic's Leadership page, as retrieved on 15 October 2026, lists Tom Brown as Co-Founder and Chief Compute Officer.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Shows the listing on the retrieval date only; when he took the title is not established.","evidenceIds":["lg-ev-ant-brown-chief-compute-2026"],"attributes":{"role":"Chief Compute Officer"}},{"id":"lg-cn-wang-shaolan-zhipu-cofounder","statement":"Zhipu’s Hong Kong prospectus of 30 December 2025 describes Wang Shaolan as a co-founder of the company.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A retrospective co-founder designation in the company’s own filing; it does not establish the person’s role at founding or a continuing role. The prospectus names six co-founders; the dataset does not establish that the list is exhaustive.","evidenceIds":["lg-ev-cn-wang-shaolan-zhipu-cofounder"]},{"id":"ap-seth-g-benzell-benzell-et-al-2015-robots-are-us","statement":"“Robots Are Us: Some Economics of Human Replacement” (2015-02) credits Seth G. Benzell as one of its 4 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-seth-g-benzell-benzell-et-al-2015-robots-are-us"]},{"id":"ap-daniel-rock-brynjolfsson-rock-syverson-2017-productivity-paradox","statement":"“Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics” (2017-11) credits Daniel Rock as one of its 3 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-daniel-rock-brynjolfsson-rock-syverson-2017-productivity-paradox"]},{"id":"cmp-rsp-2023-d1-other","statement":"RSP · 2023 states a provision on the comparison question “Other named domains”, in a weaker form than the question’s criterion.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Stated, partly: the passage has what the question’s criterion asks for, but in a weaker form than the criterion’s own words. Behaviour contrary to designers’ intent is named inside the “Autonomy and replication” risk source, not as a separate tracked domain. Two risk sources are named (misuse; autonomy and replication). Undermining of safety measures appears in the speculative ASL-4 section (PDF p. 15). 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-2023-d1-other-1"],"attributes":{"provision":{"kind":"risk_domain","key":"comparison-d1-other","facet":{"id":"D1.other","status":"partly"}}}},{"id":"cmp-rsp-v3-0-d7-external","statement":"RSP · v3.0 states a provision on the comparison question “Public or external reporting”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Also: the Frontier Safety Roadmap is published in redacted form (§2); external reviewers are asked to publish commentary within 30 days (§3.6.2); in the novel chemical/biological weapons row (PDF p. 7), the company column says threat-pathway content and policy recommendations will be shared with policymakers; where a noncompliance report is substantiated and involves a material safety risk, Anthropic “may” give public notice (§4). 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-v3-0-d7-external-1"],"attributes":{"provision":{"kind":"reporting","key":"comparison-d7-external","facet":{"id":"D7.external","status":"stated"}}}},{"id":"fl-cn-genai-art19-explain-training-data","statement":"Article 19 of the Interim Measures requires providers to cooperate with supervision and inspection by the competent authorities and, as required, to explain the source, scale and types of training data, the annotation rules and the mechanisms of their algorithms, and to provide necessary technical and data support.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A duty owed on inspection, not a duty to publish. Whether any named developer or service falls within the category is set aside.","evidenceIds":["evidence-fl-cn-genai-art19-explain-training-data-1"],"attributes":{"provision":{"kind":"required_action","key":"cn-genai-art19"}}},{"id":"fl-cn-labelling-explicit-duty","statement":"Article 4 of the Labelling Measures requires service providers whose services fall under Article 17(1) of the Deep Synthesis Provisions to add explicit labels to generated text, audio, images, video and virtual scenes, and to ensure that downloaded, copied or exported files contain the required explicit labels.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The category is defined by cross-reference to the Deep Synthesis Provisions, which were not retrieved. Whether any named developer or service falls within the category is set aside.","evidenceIds":["evidence-fl-cn-labelling-explicit-duty-1","evidence-fl-cn-labelling-explicit-duty-2"],"attributes":{"provision":{"kind":"required_action","key":"cn-labelling-art4"}}},{"id":"cpo-orb-oc2-ascend-2024","statement":"The ASCEND results release (June 2024) states that space data centres would not require water to cool them.","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 for earlier texts; not a claim that no earlier text exists. It states the water point only; it says nothing of radiators or cooling energy. The study report was not read.","evidenceIds":["evidence-cpo-orb-oc2-ascend-2024-1"],"attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction"}}},{"id":"ab-ev-vct-benchmark-result","statement":"The Virology Capabilities Test paper states that expert virologists with internet access scored an average of 22.1% on the benchmark's questions within their own sub-areas of expertise, while the most performant model it tested, OpenAI's o3, reached 43.8% accuracy and outperformed 94% of those experts within their sub-areas.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A benchmark result as the paper states it, for the models and expert comparison group it describes at that date. Scores on this benchmark are not a measure of real-world capability or risk, and nothing is generalised to other models or later versions. Only the abstract page was read, so the figures are the paper's own summary. The figures and the comparison are recorded; no question, subject matter or procedure is recorded. Paraphrase is used because the abstract page renders the numbers in LaTeX markup.","evidenceIds":["evidence-ab-ev-vct-benchmark-result-1"],"attributes":{"evaluatedSystem":"The models the paper tested, of which it names OpenAI's o3 as the most performant","evaluation":{"what":"Performance on the benchmark's multimodal questions, compared with expert virologists in their own sub-areas","how":"A 322-question multimodal benchmark scored against a human expert baseline with internet access","result":"Expert average of 22.1% within sub-areas of expertise; the most performant model reached 43.8% accuracy, outperforming 94% of experts within their sub-areas","caveats":"The paper states that the capability measured is inherently dual-use and proposes that it be taken into existing frameworks for dual-use technologies; it does not state that benchmark performance measures real-world capability."}}}],"sources":[{"id":"source-rsp-1-0","title":"Responsible Scaling Policy — v1.0","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":"2023-09-19","version":"Responsible Scaling Policy v1.0; effective date 2023-09-19.","versionDate":"2023-09-19","canonicalUrl":"https://www-cdn.anthropic.com/1adf000c8f675958c2ee23805d91aaade1cd4613/responsible-scaling-policy.pdf","evidenceUrl":"https://www-cdn.anthropic.com/1adf000c8f675958c2ee23805d91aaade1cd4613/responsible-scaling-policy.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20230919161730/https://www-files.anthropic.com/production/files/responsible-scaling-policy-1.0.pdf","archivedAt":"2023-09-19","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v1.0.","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-3-0","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-anthropic-aws-nov-2024","title":"Powering the next generation of AI development with AWS","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":"2024-11-22","version":"Dated announcement inspected 2026-09-19; a historical company statement, not independent verification of implementation.","canonicalUrl":"https://www.anthropic.com/news/anthropic-amazon-trainium","evidenceUrl":"https://www.anthropic.com/news/anthropic-amazon-trainium","language":"en","archivedUrl":"https://web.archive.org/web/20260908143654/https://www.anthropic.com/news/anthropic-amazon-trainium","archivedAt":"2026-09-08","accessedAt":"2026-09-19","reuse":{"attribution":"Anthropic, Powering the next generation of AI development with AWS (2024-11-22).","licence":"No open redistribution licence asserted","note":"Bibliographic metadata and editorial paraphrases only; no full source redistributed."}},{"id":"source-human-preferences-paper-v1","title":"Deep reinforcement learning from human preferences","authors":["Paul Christiano","Jan Leike","Tom B. Brown","Miljan Martic","Shane Legg","Dario Amodei"],"publisher":"arXiv","publicationDate":"2017-06-12","version":"arXiv:1706.03741v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2017-06-12","canonicalUrl":"https://arxiv.org/abs/1706.03741","evidenceUrl":"https://arxiv.org/pdf/1706.03741v1","language":"en","accessedAt":"2026-09-20","reuse":{"attribution":"Paul Christiano; Jan Leike; Tom B. Brown; Miljan Martic; Shane Legg; Dario Amodei, Deep reinforcement learning from human preferences (2017-06-12).","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-weak-strong-paper-v1","title":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision","authors":["Collin Burns","Pavel Izmailov","Jan Hendrik Kirchner","Bowen Baker","Leo Gao","Leopold Aschenbrenner","Yining Chen","Adrien Ecoffet","Manas Joglekar","Jan Leike","Ilya Sutskever","Jeff Wu"],"publisher":"arXiv","publicationDate":"2023-12-14","version":"arXiv:2312.09390v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2023-12-14","canonicalUrl":"https://arxiv.org/abs/2312.09390","evidenceUrl":"https://arxiv.org/pdf/2312.09390v1","language":"en","accessedAt":"2026-09-25","reuse":{"attribution":"Collin Burns; Pavel Izmailov; Jan Hendrik Kirchner; Bowen Baker; Leo Gao; Leopold Aschenbrenner; Yining Chen; Adrien Ecoffet; Manas Joglekar; Jan Leike; Ilya Sutskever; Jeff Wu, Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata and brief attributed paraphrases only; source publications and personal contact details are not redistributed."}},{"id":"source-atlas-gym","title":"OpenAI Gym Beta","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":"2016-04-27","version":"Dated publisher page inspected 2026-09-28; current HTML, not an authenticated historical capture.","canonicalUrl":"https://openai.com/index/openai-gym-beta/","evidenceUrl":"https://openai.com/index/openai-gym-beta/","language":"en","archivedUrl":"https://web.archive.org/web/20260928012330/https://openai.com/index/openai-gym-beta/","archivedAt":"2026-09-28","accessedAt":"2026-09-28","reuse":{"attribution":"OpenAI. OpenAI Gym Beta.","licence":"No redistribution licence asserted.","note":"Short attributed paraphrases only; original publication is linked, not redistributed."}},{"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-faf-announcement","title":"Our Approach to Frontier AI","authors":["Meta"],"publisher":"Meta","publicationDate":"2025-02-03","version":"Dated newsroom post; the live page also shows an update date of April 20, 2026. Quotations are checked against the Internet Archive capture of 2025-02-03 and the live page.","canonicalUrl":"https://about.fb.com/news/2025/02/meta-approach-frontier-ai/","evidenceUrl":"https://about.fb.com/news/2025/02/meta-approach-frontier-ai/","language":"en","archivedUrl":"https://web.archive.org/web/20250203200742/https://about.fb.com/news/2025/02/meta-approach-frontier-ai/","archivedAt":"2025-02-03","accessedAt":"2026-10-03","reuse":{"attribution":"Meta, Our Approach to Frontier AI (2025-02-03).","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-anthropic-coi-2023-09-21","title":"Amended and Restated Certificate of Incorporation of Anthropic, PBC (filed 2023-09-21)","authors":["Anthropic, PBC"],"publisher":"Anthropic","publicationDate":"2023-09-21","version":"Filed 2023-09-21 according to the Division of Corporations stamp on page 1 (file number 4860621); 24-page scan. Delaware-stamped scan published by AI Lab Watch in a shared Google Drive folder (linked from its post of 12 June 2024, which says the filings were obtained from the State of Delaware); downloaded directly on 2026-10-04 and read from page images, as the PDF has no text layer. The Internet Archive holds no capture of this file and no new capture was requested. Authenticity rests on the filing stamp and the consistent recitals of later certificates; no copy was obtained from the Division of Corporations. Publisher as printed on the inspected expression: Anthropic, PBC (filed with the Delaware Division of Corporations).","canonicalUrl":"https://drive.google.com/file/d/1keKQ1bSWMICL9VbYa2llQRXO0tBc0YKd/view","evidenceUrl":"https://drive.google.com/file/d/1keKQ1bSWMICL9VbYa2llQRXO0tBc0YKd/view","language":"en","archivedUrl":"https://web.archive.org/web/20261029125309/https://drive.usercontent.google.com/download?id=1keKQ1bSWMICL9VbYa2llQRXO0tBc0YKd&export=download&authuser=0","archivedAt":"2026-10-29","accessedAt":"2026-10-04","reuse":{"attribution":"Anthropic, PBC, amended and restated certificate of incorporation (filed 2023-09-21), copy published by AI Lab Watch.","licence":"Public 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-anthropic-coi-2024-01-29","title":"Amended and Restated Certificate of Incorporation of Anthropic, PBC (filed 2024-01-29)","authors":["Anthropic, PBC"],"publisher":"Anthropic","publicationDate":"2024-01-29","version":"Filed 2024-01-29 according to the Division of Corporations stamp on page 1 (file number 4860621); 26-page scan. Delaware-stamped scan published by AI Lab Watch in a shared Google Drive folder (linked from its post of 12 June 2024, which says the filings were obtained from the State of Delaware); downloaded directly on 2026-10-04 and read from page images, as the PDF has no text layer. The Internet Archive holds no capture of this file and no new capture was requested. Authenticity rests on the filing stamp and the consistent recitals of later certificates; no copy was obtained from the Division of Corporations. Publisher as printed on the inspected expression: Anthropic, PBC (filed with the Delaware Division of Corporations).","canonicalUrl":"https://drive.google.com/file/d/1TDPOkwshy__4eRVofr5LDyD_osoKusve/view","evidenceUrl":"https://drive.google.com/file/d/1TDPOkwshy__4eRVofr5LDyD_osoKusve/view","language":"en","archivedUrl":"https://web.archive.org/web/20261029125604/https://drive.usercontent.google.com/download?id=1TDPOkwshy__4eRVofr5LDyD_osoKusve&export=download","archivedAt":"2026-10-29","accessedAt":"2026-10-04","reuse":{"attribution":"Anthropic, PBC, amended and restated certificate of incorporation (filed 2024-01-29), copy published by AI Lab Watch.","licence":"Public 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-wiener-1960-science","title":"Some Moral and Technical Consequences of Automation","authors":["Norbert Wiener"],"publisher":"Science (American Association for the Advancement of Science)","publicationDate":"1960-05-06","version":"Scan of the original Science pages in a later JSTOR download, hosted on a Pennsylvania State University course page; text layer read and pp. 1357–1358 compared with the rendered pages by the pre-1996 candidate search.","canonicalUrl":"https://doi.org/10.1126/science.131.3410.1355","evidenceUrl":"https://faculty.ist.psu.edu/vhonavar/Courses/ai100/Wiener.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20241222023645/https://faculty.ist.psu.edu/vhonavar/Courses/ai100/Wiener.pdf","archivedAt":"2024-12-22","accessedAt":"2026-10-06","reuse":{"attribution":"Norbert Wiener, \"Some Moral and Technical Consequences of Automation\", Science 131 (1960), 1355–1358.","licence":"No open redistribution licence asserted; the inspected copy is a university course file.","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-moravec-1989-takeover-file","title":"Human Culture—A Genetic Takeover Underway","authors":["Hans Moravec"],"publisher":"Hans Moravec","publicationDate":"1989","version":"Author-hosted LaTeX file dated July 1988, retrieved 2026-10-06 (hash matched the pre-1996 candidate search); identical to the capture of 2001-03-06. Author’s project archive, Carnegie Mellon University. Date: the file’s header gives “July 1988” (the file and the capture alike); the publication date, 1989, is the year of the printed chapter in the author’s publication list. No 1989 printing was inspected.","versionDate":"1988-07","canonicalUrl":"https://frc.ri.cmu.edu/~hpm/project.archive/general.articles/1988/takeover.ltx","evidenceUrl":"https://frc.ri.cmu.edu/~hpm/project.archive/general.articles/1988/takeover.ltx","language":"en","archivedUrl":"https://web.archive.org/web/20010306111246/http://www.frc.ri.cmu.edu:80/~hpm/project.archive/general.articles/1988/takeover.ltx","archivedAt":"2001-03-06","accessedAt":"2026-10-06","reuse":{"attribution":"Hans Moravec, \"Human Culture—A Genetic Takeover Underway\", in Artificial Life (1989).","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-clancy-besiroglu-2023-great-inflection","title":"The Great Inflection? A Debate About AI and Explosive Growth","authors":["Matt Clancy","Tamay Besiroglu"],"publisher":"Asterisk magazine, issue 03","publicationDate":"2023-06","version":"Wayback capture of asteriskmag.com/issues/03/the-great-inflection-a-debate-about-ai-and-explosive-growth. Capture retrieved 2026-10-08 (SHA-256 e8b29f97be40d49c282cac8d279cc008644b822a4607533999157375d01753d1). Published version: Clancy, M., & Besiroglu, T. (2023, June). The great inflection? A debate about AI and explosive growth. Asterisk, (3). Magazine debate.","canonicalUrl":"https://asteriskmag.com/issues/03/the-great-inflection-a-debate-about-ai-and-explosive-growth","evidenceUrl":"https://web.archive.org/web/20231223213358/https://asteriskmag.com/issues/03/the-great-inflection-a-debate-about-ai-and-explosive-growth","language":"en","archivedUrl":"https://web.archive.org/web/20231223213358/https://asteriskmag.com/issues/03/the-great-inflection-a-debate-about-ai-and-explosive-growth","archivedAt":"2023-12-23","accessedAt":"2026-10-08","reuse":{"attribution":"Matt Clancy, Tamay Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth”, Asterisk magazine, issue 03, 2023-06.","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-benzell-et-al-2015-robots-are-us","title":"Robots Are Us: Some Economics of Human Replacement","authors":["Seth G. Benzell","Laurence J. Kotlikoff","Guillermo LaGarda","Jeffrey D. Sachs"],"publisher":"NBER Working Paper 20941","publicationDate":"2015-02","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 362172ebc8a725bfe76675a836f565736278dc12c86db96b5e67944451b0ae60. The February 2015 wording was not retrieved; the retrieved PDF is the October 2018 revision. Benzell, S. G., Kotlikoff, L. J., LaGarda, G., & Sachs, J. D. (2015). Robots are us: Some economics of human replacement (NBER Working Paper No. 20941; February 2015 version). National Bureau of Economic Research. https://doi.org/10.3386/w20941","canonicalUrl":"https://www.nber.org/system/files/working_papers/w20941/w20941.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w20941/w20941.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Seth G. Benzell, Laurence J. Kotlikoff, Guillermo LaGarda, Jeffrey D. Sachs, “Robots Are Us: Some Economics of Human Replacement”, NBER Working Paper 20941, 2015-02.","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-brynjolfsson-rock-syverson-2017-productivity-paradox","title":"Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics","authors":["Erik Brynjolfsson","Daniel Rock","Chad Syverson"],"publisher":"NBER Working Paper 24001","publicationDate":"2017-11","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 7c34b5a8edac034b1ec781568cfb8e33b2d37e30705ce24424c2fc1a1f837992.","canonicalUrl":"https://www.nber.org/system/files/working_papers/w24001/w24001.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w24001/w24001.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Erik Brynjolfsson, Daniel Rock, Chad Syverson, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics”, NBER Working Paper 24001, 2017-11.","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-ant-leadership-2026","title":"Leadership at Anthropic (web page, undated; retrieved 2026-10-15)","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Undated page, retrieved on 2026-10-15 (SHA-256 a2bc8182bd18aff065d4bce72994c832816f98e4f099d956173e849b7baa789c) directly from the publisher. Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.anthropic.com/company/leadership","evidenceUrl":"https://www.anthropic.com/company/leadership","language":"en","archivedUrl":"https://web.archive.org/web/20261015210858/https://www.anthropic.com/company/leadership","archivedAt":"2026-10-15","accessedAt":"2026-10-15","reuse":{"attribution":"Anthropic, “Leadership at Anthropic (web page, undated; retrieved 2026-10-15)”.","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":"lg-src-cn-minimax-prospectus","title":"MiniMax Group Inc. — Global Offering","authors":["MiniMax Group Inc."],"publisher":"MiniMax","publicationDate":"2025-12-31","version":"Dated document, retrieved on 2026-10-15 (SHA-256 7853021e46da5714242f8d27984498cbfef3e260d5a7cbb8d4e6d62d83ab9df3) directly from the publisher. Publisher as printed on the inspected expression: MiniMax Group Inc. (via HKEXnews).","canonicalUrl":"https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1231/2025123100025.pdf","evidenceUrl":"https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1231/2025123100025.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260930092156/https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1231/2025123100025.pdf","archivedAt":"2026-09-30","accessedAt":"2026-10-15","reuse":{"attribution":"MiniMax Group Inc., “MiniMax Group Inc. — Global Offering”, 2025-12-31.","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-cn-zhipu-prospectus-en","title":"Knowledge Atlas Technology Joint Stock Company Limited (北京智譜華章科技股份有限公司) — Global Offering","authors":["Knowledge Atlas Technology Joint Stock Company Limited"],"publisher":"Knowledge Atlas Technology Joint Stock Company Limited (via HKEXnews)","publicationDate":"2025-12-30","version":"Dated document, retrieved on 2026-10-15 (SHA-256 a6598ce3f2f13ac78a9d02a85314dcd349314d2ee1215015e41bb1284e608f4e) directly from the publisher.","canonicalUrl":"https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1230/2025123000017.pdf","evidenceUrl":"https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1230/2025123000017.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20261002230645/https://www1.hkexnews.hk/listedco/listconews/sehk/2025/1230/2025123000017.pdf","archivedAt":"2026-10-02","accessedAt":"2026-10-15","reuse":{"attribution":"Knowledge Atlas Technology Joint Stock Company Limited, “Knowledge Atlas Technology Joint Stock Company Limited (北京智譜華章科技股份有限公司) — Global Offering”, 2025-12-30.","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-deal-msft-suleyman-2024","title":"Mustafa Suleyman, DeepMind and Inflection Co-founder, joins Microsoft to lead Copilot","authors":["Satya Nadella"],"publisher":"Microsoft","publicationDate":"2024-03-19","version":"Dated page, retrieved on 2026-10-15 (SHA-256 45613ab2bd14157eef4a4ebe9eedcb47d84a36058b3ba099115c64844dc327ad) directly from the publisher.","canonicalUrl":"https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/","evidenceUrl":"https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/","language":"en","archivedUrl":"https://web.archive.org/web/20261014225026/https://blogs.microsoft.com/blog/2024/03/19/mustafa-suleyman-deepmind-and-inflection-co-founder-joins-microsoft-to-lead-copilot/","archivedAt":"2026-10-14","accessedAt":"2026-10-15","reuse":{"attribution":"Satya Nadella, “Mustafa Suleyman, DeepMind and Inflection Co-founder, joins Microsoft to lead Copilot”, 2024-03-19.","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-gc-thepaper-baidu-2025","title":"百度新设基础模型和应用模型研发部，直接向李彦宏汇报","authors":["吴雨欣"],"publisher":"澎湃新闻 (The Paper)","publicationDate":"2025-11-25","version":"Dated page, retrieved on 2026-10-15 (SHA-256 b963571522865094aff0e67122fe9bc097cf9345675bf6245a8a5ddfa3e039bc) directly from the publisher.","canonicalUrl":"https://www.thepaper.cn/newsDetail_forward_32041037","evidenceUrl":"https://www.thepaper.cn/newsDetail_forward_32041037","language":"zh","archivedUrl":"https://web.archive.org/web/20261029004401/https://www.thepaper.cn/newsDetail_forward_32041037","archivedAt":"2026-10-29","accessedAt":"2026-10-15","reuse":{"attribution":"吴雨欣, “百度新设基础模型和应用模型研发部，直接向李彦宏汇报”, 2025-11-25.","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":"lg-src-gw-sifted-h-2024","title":"Three cofounders leave French AI startup H just three months after raising $220m seed","authors":["Daphné Leprince-Ringuet"],"publisher":"Sifted","publicationDate":"2024-08-24","version":"Dated page, retrieved on 2026-10-15 (SHA-256 c01ab67faa496e9be6d7e4d4df5f83d688ff835ef0adf7d1b7d7a9f9f6d1cf21) 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://sifted.eu/articles/three-cofounders-leave-h-news","evidenceUrl":"https://sifted.eu/articles/three-cofounders-leave-h-news","language":"en","archivedUrl":"https://web.archive.org/web/20260416133943/https://sifted.eu/articles/three-cofounders-leave-h-news","archivedAt":"2026-04-16","accessedAt":"2026-10-15","reuse":{"attribution":"Daphné Leprince-Ringuet, “Three cofounders leave French AI startup H just three months after raising $220m seed”, 2024-08-24.","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":"lg-src-oai-ssc-2024","title":"OpenAI Board Forms Safety and Security Committee","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":"2024-05-28","version":"Dated page, retrieved on 2026-10-15 (SHA-256 34a351ce00b93ca65fdcd73f59225314098c44da2c1865630989fa68cd213a52) 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. The inspected text carries an update of 18 June 2024 above the original announcement.","canonicalUrl":"https://openai.com/index/openai-board-forms-safety-and-security-committee/","evidenceUrl":"https://openai.com/index/openai-board-forms-safety-and-security-committee/","language":"en","archivedUrl":"https://web.archive.org/web/20260713143403/https://openai.com/index/openai-board-forms-safety-and-security-committee/","archivedAt":"2026-07-13","accessedAt":"2026-10-15","reuse":{"attribution":"OpenAI, “OpenAI Board Forms Safety and Security Committee”, 2024-05-28.","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-oai-weng-x-2024","title":"Post of 8 November 2024 on leaving OpenAI","authors":["Lilian Weng"],"publisher":"X","publicationDate":"2024-11-08","version":"Dated post, retrieved on 2026-10-15 (SHA-256 554fd15f767467c91ed821afcf9bc68424bef151d4c630149ccac82e60fa51cb) through X’s own syndication endpoint for the post’s text, which truncates long notes. Read through X’s syndication endpoint (text only). Publication date is the post’s UTC timestamp.","canonicalUrl":"https://x.com/lilianweng/status/1855031273690984623","evidenceUrl":"https://x.com/lilianweng/status/1855031273690984623","language":"en","archivedUrl":"https://web.archive.org/web/20241108233455/https://twitter.com/lilianweng/status/1855031273690984623","archivedAt":"2024-11-08","accessedAt":"2026-10-15","reuse":{"attribution":"Lilian Weng, “Post of 8 November 2024 on leaving OpenAI”, 2024-11-08.","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-cn-generative-ai-measures-2023-govcn","title":"生成式人工智能服务管理暂行办法 (Interim Measures for the Management of Generative Artificial Intelligence Services), gov.cn State Council policy library copy","authors":["国家互联网信息办公室 (Cyberspace Administration of China) and six other departments"],"publisher":"中国政府网 (www.gov.cn), 国务院政策文件库","publicationDate":"2023-07-10","version":"Retrieved on 2026-10-27 (SHA-256 00bd290a9d14e7354b7131d30255daae26ce7ff73588483cbeb2030f56878277). A second posting of the text of the dataset's source for the Measures (the CAC's posting of 13 July 2023): the gov.cn State Council policy library page, which gives its source as 国家网信办网站 and its document date (成文日期) as 2023-07-10. The page body repeats the text twice (a print copy). The two postings were not compared word for word; every passage cited from this source was matched in this copy. The quotation limit is applied to this source and the CAC posting together. The English in parentheses in the title is a machine translation.","canonicalUrl":"https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm","evidenceUrl":"https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm","language":"zh-Hans","accessedAt":"2026-10-27","reuse":{"attribution":"国家互联网信息办公室 (Cyberspace Administration of China) and six other departments, “生成式人工智能服务管理暂行办法 (Interim Measures for the Management of Generative Artificial Intelligence Services), gov.cn State Council policy library copy”, 2023-07-10.","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":"source-cn-labelling-measures-2025","title":"关于印发《人工智能生成合成内容标识办法》的通知 (Notice on issuing the Measures for Labelling AI-Generated Synthetic Content), 国信办通字〔2025〕2号","authors":["国家互联网信息办公室 (Cyberspace Administration of China)","工业和信息化部 (Ministry of Industry and Information Technology)","公安部 (Ministry of Public Security)","国家广播电视总局 (National Radio and Television Administration)"],"publisher":"中国政府网 (www.gov.cn), 国务院政策文件库","publicationDate":"2025-03-07","version":"Retrieved on 2026-10-27 (SHA-256 0c0d1d32be9f69843088b0ad2664734b401485166bd04bdfdf35d42d9b77b84a). gov.cn policy-library page (source given as 国家网信办网站; 成文日期 2025-03-07), retrieved 2026-10-27. The English in parentheses in the title is a machine translation.","canonicalUrl":"https://www.gov.cn/zhengce/zhengceku/202503/content_7014286.htm","evidenceUrl":"https://www.gov.cn/zhengce/zhengceku/202503/content_7014286.htm","language":"zh-Hans","accessedAt":"2026-10-27","reuse":{"attribution":"国家互联网信息办公室 (Cyberspace Administration of China), 工业和信息化部 (Ministry of Industry and Information Technology), 公安部 (Ministry of Public Security), 国家广播电视总局 (National Radio and Television Administration), “关于印发《人工智能生成合成内容标识办法》的通知 (Notice on issuing the Measures for Labelling AI-Generated Synthetic Content), 国信办通字〔2025〕2号”, 2025-03-07.","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-tas-ascend-results","title":"Thales Alenia Space reveals results of ASCEND feasibility study on space data centers","authors":["Thales Alenia Space"],"publisher":"Thales Alenia Space","publicationDate":"2024-06-27","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 0f7cd327680b3e61754037f19421d3940edb0c41c780b0494e7a8aca98bb54b0). Press release as served on 2026-10-28; dateline Cannes, June 27, 2024. The ASCEND study report itself was not retrieved. Retrieved again at admission on 2026-10-30 (SHA-256 0f7cd327680b3e61754037f19421d3940edb0c41c780b0494e7a8aca98bb54b0, the same bytes). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.thalesaleniaspace.com/en/press-releases/thales-alenia-space-reveals-results-ascend-feasibility-study-space-data-centers-0","evidenceUrl":"https://www.thalesaleniaspace.com/en/press-releases/thales-alenia-space-reveals-results-ascend-feasibility-study-space-data-centers-0","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Thales Alenia Space, “Thales Alenia Space reveals results of ASCEND feasibility study on space data centers”, 2024-06-27.","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-vct-abs","title":"arXiv abstract page for arXiv:2504.16137 (Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark)","authors":["arXiv (Cornell University)"],"publisher":"arXiv (Cornell University)","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 3ad6043cc84942ab180140c67ad4e128b37899ee1e4fc957d8f80d93a80d4898). Abstract page as served on 2026-10-28, listing v1 of 21 April 2025 and v2 of 29 April 2025 (the current version). Only the abstract page was read; the full paper was not retrieved, so no record rests on its body. Retrieved again at admission on 2026-10-30 (SHA-256 cc6f6960f2503c7789a77800e49cac2de3af708305ca7f232a5322507a12bc80). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://arxiv.org/abs/2504.16137","evidenceUrl":"https://arxiv.org/abs/2504.16137","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"arXiv (Cornell University), “arXiv abstract page for arXiv:2504.16137 (Virology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark)”.","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-amazon-anthropic-new-2024","sourceRole":"primary","support":"supports","locator":"Opening paragraph","passage":"The expansion included a new $4 billion investment; the total would become $8 billion.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-19","note":"Exact cited source passage inspected by a machine agent. Company statements are not independent proof of implementation. Human source review pending."},{"id":"evidence-lineage-jan-leike-preferences","sourceRole":"primary","support":"supports","locator":"v1 PDF p. 1, author block","passage":"The named person appears in the author list.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-20","note":"Relevant source passage inspected by machine. Primary for what this document states; no independent replication or human review. Existing source access dates are retained; this evidence check has its own date."},{"id":"evidence-lineage-ii-weak-cites-critique","sourceRole":"primary","support":"supports","locator":"PDF §2, p. 5, Scalable oversight; bibliography p. 25","passage":"The related-work paragraph cites this work in its account of scalable oversight.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-25","note":"Direct source inspection; machine check only. Work-level citation; the cited revision is unspecified. Does not establish agreement, causal influence, implementation or institutional adoption. The paragraph contrasts improving supervision with generalizing beyond it; no method-use edge is justified."},{"id":"evidence-atlas-gym","sourceRole":"primary","support":"supports","locator":"Opening two paragraphs","passage":"The release describes environments for developing and comparing RL algorithms.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-28","note":"Original source passage checked by the machine agent that proposed the record. No independent machine audit or human source inspection is recorded for this increment."},{"id":"evidence-issued-meta-faf-v1-1","sourceRole":"primary","support":"supports","locator":"Takeaways, first item (page dated February 3, 2025)","passage":"Today, we’re sharing our Frontier AI Framework, which outlines our consideration of risk in our model-release decisions, in line with the commitment we made at last year’s global AI Seoul Summit.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-03","note":"Verbatim excerpt checked by an automated research assistant on 2026-10-03 against the Internet Archive capture of 2025-02-03 and the live page. Human editorial review is pending."},{"id":"evidence-issued-meta-faf-v1-1-2","sourceRole":"primary","support":"supports","locator":"PDF p. 44, Appendix II – Change log, first entry","passage":"Meta’s later change log dates the Frontier AI Framework’s initial version February 3, 2025.","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-prov-meta-aasf-v2-0-high-deployment-measure","sourceRole":"primary","support":"supports","locator":"PDF p. 15 (printed p. 15), §3.3 Risk Thresholds, Table 1, High row, Measures column, under the label 'Deploy with mitigations'","passage":"Proceed with deployment of the Frontier AI only if sufficient mitigations are defined, implemented and validated to reduce risk to that of a moderate or lower model.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-03","note":"Verbatim excerpt checked by an automated research assistant on 2026-10-03 against the file served today, byte-identical to the Internet Archive capture of 2026-04-08. Extraction note: The identical sentence also appears in the Critical row's Measures cell on the same page. Human editorial review is pending."},{"id":"evidence-cs-openai-990-2019-gp-llc-controls-lp-1","sourceRole":"primary","support":"supports","locator":"Schedule R, Part III, line (1), columns (a), (d) and (j)","passage":"Part III lists OpenAI LP (EIN 83-1960637) as a related organisation taxable as a partnership, gives OpenAI GP LLC as its direct controlling entity and answers “Yes” in the general-or-managing-partner column.","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-provchange-anthropic-coi-2024-01-29-class-t-phase-in-1","sourceRole":"primary","support":"supports","locator":"Article VIII(B)6(iv) and (vi) (page 22)","passage":"The 2024 certificate sets the funding-based triggers at eight months and four months following the board’s determination that $6 billion and $5 billion have been raised.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-04","note":"Editorial paraphrase of the page image, re-checked on 2026-10-04 by the integrating session against the scan. Added by the integrating session in week 4c; not part of the week 4b research pass. Human editorial review is pending."},{"id":"evidence-provchange-anthropic-coi-2024-01-29-class-t-phase-in-2","sourceRole":"primary","support":"supports","locator":"Article VIII(B)6(iv) and (vi) (page 21)","passage":"The September 2023 certificate sets the funding-based triggers at the time of the board’s determination itself.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-04","note":"Editorial paraphrase of the page image, re-checked on 2026-10-04 by the integrating session against the scan. Added by the integrating session in week 4c; not part of the week 4b research pass. Human editorial review is pending."},{"id":"evidence-ts-a4-wiener-1960-context","sourceRole":"primary","support":"supports","locator":"Printed p. 1357, right column; pp. 1355 and 1358 for time scales","passage":"If a machine’s information processing outpaces the operator, the operator may learn that it should be stopped only after the damage is done.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-06","note":"Paraphrase from the pre-1996 candidate search."},{"id":"evidence-ts-a4-moravec-1989-context","sourceRole":"primary","support":"supports","locator":"Opening paragraphs","passage":"Unfettered by biology, machines are overtaking people.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-06","note":"Paraphrase checked against the retrieved file."},{"id":"evidence-ec-g03-f","sourceRole":"primary","support":"supports","locator":"Tamay Besiroglu's turn beginning \"In my view, accelerating growth is probably not decoupled\"","passage":"In this sense, economic acceleration roughly on the order we’re considering for this debate is perhaps actually something of a historical norm.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage G03-f (research/economics-growth). The quotation (22 words) matched the checked expression (asterisk-debate) 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. On 2026-10-30, by the maintainer’s decision of that day, the quotation was corrected to the archive capture of 2023-12-23 (quotation marks and apostrophes as in the capture), which a cross-family archive check read on 2026-10-28; the proposition is unchanged."},{"id":"lg-ev-cn-yun-yeyi-sensetime-2017-2022","sourceRole":"primary","support":"supports","locator":"Directors and Senior Management, “Executive Directors”, biography of Yun Yeyi, printed page 352","passage":"she served various positions in SenseTime Group Inc. … from September 2017 … Ms. Yun then served as its director of innovative business from January 2021 to January 2022","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (29 words) matched the retrieved text on 2026-10-15 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. Elided quotation: the ellipses mark omitted words. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-deal-simonyan-microsoft-ai-chief-scientist-2024","sourceRole":"primary","support":"supports","locator":"Paragraph 4","passage":"Karén is joining this group as Chief Scientist, reporting to Mustafa.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (11 words) matched the retrieved text on 2026-10-15 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. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-gw-perolat-h-company-departure-2024-a","sourceRole":"secondary","support":"supports","locator":"Standfirst","passage":"Karl Tuyls, Daan Wierstra and Julien Perolat are leaving the company","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (11 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. 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-gw-perolat-h-company-departure-2024-b","sourceRole":"secondary","support":"supports","locator":"Paragraph 1","passage":"Three of the five cofounders behind French AI startup H Company have left just three months after raising a $220m seed round","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (22 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. 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-oai-weng-departure-2024","sourceRole":"primary","support":"supports","locator":"Post text, first sentence","passage":"Weng writes that after almost seven years at OpenAI she has decided to leave.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “decide to leave” 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":"lg-ev-oai-weng-departure-2024-title","sourceRole":"primary","support":"context","locator":"Paragraph listing the technical and policy experts on the committee (28 May 2024)","passage":"OpenAI’s committee announcement of 28 May 2024 identifies Weng as Head of Safety Systems.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Weng (Head of Safety Systems)” 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. No human editorial review.","retrievalRoute":"reader_relay"},{"id":"lg-ev-gc-jia-lei-baidu-application-model-2025","sourceRole":"secondary","support":"supports","locator":"Second paragraph","passage":"The report states that Baidu set up a new Application Model R&D Department, responsible for tuning and exploring the specialised models that business applications need, headed by Jia Lei.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “由贾磊负责” matched the retrieved text on 2026-10-15 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. The paraphrase is Selekoa’s; only the anchor is the source’s wording. Paraphrase of a passage not in English; no human translation review. A cross-family machine check (2026-10-16/17) re-read the record’s sources and agreed with the record. Shortened at admission to keep the quotations from this source within the per-document limit; the shortened wording is part of the checked passage. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-ant-brown-chief-compute-2026","sourceRole":"primary","support":"supports","locator":"Section “Co-Founders”, Tom Brown entry","passage":"Tom Brown's entry gives his title as Co-Founder and Chief Compute Officer.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Chief Compute Officer” matched the retrieved text on 2026-10-15 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. The paraphrase is Selekoa’s; only the anchor is the source’s wording. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-cn-wang-shaolan-zhipu-cofounder","sourceRole":"primary","support":"supports","locator":"History, Development and Corporate Structure, lock-up table note (1), printed page 142","passage":"Mr. Wang Shaolan (our co-founder)","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (5 words) matched the retrieved text on 2026-10-15 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. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ap-seth-g-benzell-benzell-et-al-2015-robots-are-us","sourceRole":"primary","support":"supports","locator":"PDF text layer, first two pages: author name on the title page or first page","passage":"Seth G. Benzell","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-daniel-rock-brynjolfsson-rock-syverson-2017-productivity-paradox","sourceRole":"primary","support":"supports","locator":"PDF text layer, first two pages: author name on the title page or first page","passage":"Daniel Rock","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. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"ev-cmp-rsp-2023-d1-other-1","sourceRole":"primary","support":"supports","locator":"PDF p. 3 (printed p. 3), Framework › 'Sources of Catastrophic Risk', second bullet ('Autonomy and replication'), first sentence","passage":"Under the “Autonomy and replication” risk source, the edition says scaled systems may become capable of greater autonomy that lets them proliferate and, because current methods for steering them are imperfect, could potentially act against what their designers or users intend.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-27","note":"Explicit paraphrase, written on 2026-10-27 by an automated research assistant after re-reading the passage in the publisher copy, whose SHA-256 matches the copy retrieved on 2026-10-15. The comparison quoted the passage; it is given here as a paraphrase to keep the edition within the quotation limit. Human editorial review is pending."},{"id":"ev-cmp-rsp-v3-0-d7-external-1","sourceRole":"primary","support":"supports","locator":"PDF p. 13, §3.5 Publication and Redactions, first sentence","passage":"The edition says Anthropic will make a version of its Risk Report public.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-27","note":"Explicit paraphrase, written on 2026-10-27 by an automated research assistant after re-reading the passage in the publisher copy, whose SHA-256 matches the copy retrieved on 2026-10-15. The comparison quoted the passage; it is given here as a paraphrase to keep the edition within the quotation limit. The same section says Anthropic will aim to keep redactions to a minimum. Human editorial review is pending."},{"id":"evidence-fl-cn-genai-art19-explain-training-data-1","sourceRole":"primary","support":"supports","locator":"Article 19 (第十九条), first paragraph","passage":"按要求对训练数据来源、规模、类型、标注规则、算法机制机理等予以说明","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (29 words, each CJK character counted as one) 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 www.gov.cn. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. English machine translation, not verified by a person: “as required, give explanations of the source, scale and types of training data, the annotation rules, the mechanisms of the algorithms, and so on”. A cross-family machine check (2026-10-28, item 100) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-fl-cn-labelling-explicit-duty-1","sourceRole":"primary","support":"supports","locator":"Article 4 (第四条), second paragraph","passage":"服务提供者提供生成合成内容下载、复制、导出等功能时，应当确保文件中含有满足要求的显式标识","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (41 words, each CJK character counted as one) 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 www.gov.cn. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. English machine translation, not verified by a person: “When service providers offer functions such as downloading, copying or exporting generated synthetic content, they shall ensure the files contain explicit labels meeting the requirements”. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-fl-cn-labelling-explicit-duty-2","sourceRole":"primary","support":"supports","locator":"Article 4 (第四条), first paragraph, points (1)–(6)","passage":"The opening paragraph specifies placement of explicit labels for text, audio, images, video, virtual scenes and other scenarios, for services within Article 17(1) of the Deep Synthesis Provisions.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The paraphrase is Selekoa’s, written from the text retrieved on 2026-10-27 and read directly from www.gov.cn; it is not a quotation and was not machine-matched. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc2-ascend-2024-1","sourceRole":"primary","support":"supports","locator":"paragraph 4","passage":"space data centers would not require water to cool them","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (10 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 82) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-vct-benchmark-result-1","sourceRole":"primary","support":"supports","locator":"Abstract, third and fourth sentences","passage":"The abstract states that expert virologists with access to the internet score an average of 22.1% on questions in their own sub-areas of expertise, and that the most performant model tested, OpenAI's o3, reaches 43.8% accuracy, outperforming 94% of expert virologists even within their sub-areas of specialisation.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). 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. Explicit paraphrase: the abstract page renders the percentages as LaTeX, so a verbatim quotation would not match the retrieved text. A cross-family machine check (2026-10-30, item 28) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"}],"at":{"entities":[0,1,5,21,32,45,74,92,102,103,125,157,187,190,199,200,208,209,228,229,262,270,339,503,506,532,537,539,635,679,699,712,717,729,743,793,836,855,915,1009,1095,1116,1252],"claims":[118,203,240,284,391,473,630,665,785,820,1264,1588,1681,1721,1734,1811,1904,1946,2191,2315,2486,2538,2897,2907,3081,3381],"sources":[35,39,60,92,102,126,180,181,207,223,224,262,272,483,488,491,569,582,587,600,622,685,763,770,836,837,924,1091],"evidence":[137,222,261,305,414,415,504,753,830,831,995,1031,1580,1986,2121,2188,2189,2215,2216,2330,2466,2511,2836,2962,3133,3189,3734,3751,3752,4047,4644]},"version":"1lg47t7"}