{"entities":[{"id":"openai","name":"OpenAI"},{"id":"anthropic","name":"Anthropic"},{"id":"google-brain","name":"Google Brain"},{"id":"google","name":"Google"},{"id":"rsp-v3-2","name":"Anthropic Responsible Scaling Policy — v3.2"},{"id":"faif-2025","name":"xAI Frontier Artificial Intelligence Framework — December 2025 edition"},{"id":"andrej-karpathy","name":"Andrej Karpathy"},{"id":"human-preferences-paper","name":"Deep reinforcement learning from human preferences"},{"id":"summarize-feedback-paper","name":"Learning to summarize from human feedback"},{"id":"alibaba-group","name":"Alibaba Group / 阿里巴巴集团"},{"id":"alibaba-qwen3-policy","name":"Alibaba Qwen3 open-access announcement"},{"id":"wh-voluntary-ai-commitments-2023-09","name":"Voluntary AI Commitments (September 2023)"},{"id":"gpai-code-of-practice-2025","name":"General-Purpose AI Code of Practice (final version, 10 July 2025)"},{"id":"moravec-1989-genetic-takeover","name":"Human Culture—A Genetic Takeover Underway"},{"id":"recursive-self-improvement","name":"Recursive self-improvement"},{"id":"build-superintelligence-fast","name":"Build superintelligence as fast as possible"},{"id":"developers-know-what-they-are-doing","name":"The developers know what they are doing"},{"id":"extropians-2000-09-29-yudkowsky-6073","name":"Eugene's nuclear threat"},{"id":"lw-2008-fighting-a-rearguard-action-against-the","name":"Fighting a Rearguard Action Against the Truth"},{"id":"other-constraints-prevent-explosive-growth","name":"Other constraints prevent explosive growth"},{"id":"automation-as-ordinary-technical-change","name":"Computers and AI as ordinary technical change"},{"id":"machines-cheaper-at-every-task","name":"Machines cheaper than labour at every task"},{"id":"davidson-2021-explosive-economic-growth","name":"Could Advanced AI Drive Explosive Economic Growth?"},{"id":"acemoglu-restrepo-2019-wrong-kind-of-ai","name":"The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand"},{"id":"agrawal-gans-goldfarb-2019-economics-of-ai-introduction","name":"Introduction to \"The Economics of Artificial Intelligence: An Agenda\""},{"id":"sporn-haggerty-1966-comment","name":"Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I"},{"id":"noam-shazeer","name":"Noam Shazeer"},{"id":"avi-goldfarb","name":"Avi Goldfarb"},{"id":"seoul-ministerial-statement-2024","name":"Seoul Ministerial Statement for advancing AI safety, innovation and inclusivity (22 May 2024)"},{"id":"paris-ai-action-statement-2025-web","name":"Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet (Élysée web page)"},{"id":"elysee","name":"Présidence de la République française (Élysée)"},{"id":"commonwealth-fusion-systems","name":"Commonwealth Fusion Systems (CFS)"},{"id":"sandbrink-ai-biological-misuse-2023-06","name":"Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools"},{"id":"ab-prop-design-tools-distinct","name":"Biological design tools pose risks distinct from those of language models"}],"claims":[{"id":"lineage-summaries-cites-preferences","statement":"Learning to summarize from human feedback cites Human Preferences.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An explicit work-level reference only; neither agreement nor a causal influence claim. The citing text does not identify a particular arXiv revision of the target.","evidenceIds":["evidence-lineage-summaries-cites-preferences"]},{"id":"cn-alibaba-qwen3-access","statement":"Alibaba’s Qwen3 announcement describes downloadable models and API access still to come.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The source distinguishes available downloads from forthcoming Model Studio API access. Its open-source wording is not an independent licence certification.","evidenceIds":["evidence-cn-alibaba-qwen3-access"]},{"id":"absent-rsp-v3-2-no-train-deploy-without-safeguards","statement":"No passage stating the provision “Commitment not to train or deploy without adequate safeguards” was found in RSP · v3.2 within the inspected scope.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: RSP · v2.2 states this provision. The inspected scope of RSP · v3.2 was: RSP v3.2 (effective April 29, 2026), whole text layer, PDF pp. 1–20: Introduction (pp. 3–4); §1 with the threshold table (pp. 4–10); §2–§3 Risk Reports (pp. 11–16); §4 Governance (p. 16); Appendix A (pp. 16–17); Appendix B (p. 17); Changelog (pp. 18–20). Method: Same reading, comparison and case-insensitive search terms as for v3.0, plus a word-level comparison with v3.1 (differences: LTBT-requested external review, LTBT approval of reviewers, LTBT briefings, spacing). No passage stating a commitment not to train or deploy models unless adequate safeguards are implemented, or a commitment to pause training or deployment until such safeguards are in place, was found within the operative text. Hits examined and not counted: the Introduction (p. 3), §1 (p. 4) and Appendix A scenario commitments (p. 17) as in v3.0; the Appendix A introduction sentences on cautionary action and considering pausing (p. 17), as in v3.1; the reproduced RSP v2.0 changelog entry (p. 18). 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":["evidence-absent-rsp-v3-2-no-train-deploy-without-safeguards"],"attributes":{"provision":{"kind":"required_action","key":"rsp-no-train-deploy-without-safeguards"}}},{"id":"prov-wh-2023-09-scope","statement":"The September 2023 commitments text states that, where commitments mention particular models, they apply only to generative models that are overall more powerful than the current most advanced model produced by the company making the commitment.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the text says. Whether this scope governs the seven July 2023 companies, and whether a given model fell within it, are set aside.","evidenceIds":["evidence-prov-wh-2023-09-scope-1"],"attributes":{"provision":{"kind":"threshold","key":"scope"}}},{"id":"cm-eu-code-openai","statement":"OpenAI appears on the European Commission’s list of signatories of the General-Purpose AI Code of Practice from the list’s first publication on 1 August 2025, and is still listed in the version last updated on 31 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Listing is the Commission’s statement. It is not evidence of compliance, and the list does not say which chapters each organisation signed; the Commission’s questions and answers say that opting out of a chapter loses the benefit of facilitated demonstration of compliance for it (context below). The legal effect of signing is set aside.","evidenceIds":["evidence-cm-eu-code-openai-1","evidence-cm-eu-code-openai-2","evidence-cm-eu-code-openai-3"],"attributes":{"commitment":{"stage":"listed"}}},{"id":"ts-a3-moravec-1989-context","statement":"Moravec’s chapter expects machines to manage their own design.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Self-design without the criterion’s feedback.","evidenceIds":["evidence-ts-a3-moravec-1989-context"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"No improvement of the capacity to improve is stated in the opening paragraphs.","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":"yk-y22-lw-2008-fighting-a-rearguard-action-against-the-b","statement":"The post “Fighting a Rearguard Action Against the Truth” (2008-09-24) articulates Y22 in retrospect. His 2008 account of his earlier self states that the 1996–2000 aim was speed to any superintelligence, and his next paragraph supplies the premise: the assumption that movement toward superintelligence was 'a pure good without a hint of danger'.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Premise addressed: Reaching superintelligence is good by default, so speed is what matters. Overcoming Bias post of 2008-09-24 (Coming of Age sequence), imported to LessWrong. A retrospective: the 2008 wording reports the strategy he held in 1996–2000; it is not that strategy's own wording and cannot be earliest located. The quoted paragraph states the speed strategy; the next paragraph supplies its default-good premise. Quote stops mid-sentence; it continues with a dash-separated list of methods. The rejection follows in the next paragraphs and in \"The Magnitude of His Own Folly\"; never present it as his 2008 position. That a later report cannot pair as an earlier self with a later rejection does not make the reported proposition incomplete.","evidenceIds":["evidence-yk-t22-b"],"attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","retrospective":{"from":"1996","until":"2000","note":"Reports the strategy of “Eliezer1996-2000”."}}}},{"id":"yk-y22-lw-2008-the-magnitude-of-his-own-folly-c","statement":"The post “The Magnitude of His Own Folly” (2008-09-30) contests Y22. Explicitly rejects his earlier full-speed strategy. He says his childhood technophilia broke when he understood that a failed superintelligence project could kill everyone, however good the arguments for proceeding.","evidenceIds":["evidence-yk-t22-c"]},{"id":"yk-y24-extropians-2000-09-29-yudkowsky-6080-a","statement":"The Extropians message “Re: Eugene's nuclear threat” (2000-09-29) bears on Y24. Correcting his earlier post in a thread on Eugene Leitl's suggestion of nuclear strikes on AI facilities, he says the only things he can do are to write up Friendly AI and to try to be the first to build AI.","evidenceIds":["evidence-yk-t24-e2"]},{"id":"yk-y28-extropians-2000-09-29-yudkowsky-6073-b","statement":"The Extropians message “Eugene's nuclear threat” (2000-09-29) bears on Y28. Rejecting Eugene Leitl's statement that he would nuke AI facilities facing an Armageddon-class threat, he argues that people with the experience to build an AI who say they have worked hard on Friendliness should be trusted over those who would use force.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Premise addressed: The most advanced projects are the most competent to handle safety. Argues against nuclear coercion of AI projects; his deference is conditional on builders who have devoted considerable effort to Friendliness, not to leading laboratories in general. 53 minutes later, in his reply ‘Re: Eugene's nuclear threat’ (also recorded for ‘Build it before rivals do’, Y24), he partly corrected himself: good intentions are not enough. Recorded as related, not earlier-self (cross-family check).","evidenceIds":["evidence-yk-t28-e1"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Deference conditional on builders who have worked hard on Friendliness; it does not claim that the leading projects in general can handle safety."}}},{"id":"ec-g05-davidson-2021-explosive-economic-growth-c","statement":"Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) contests G05. Responds to a list of bottlenecks (regulation, materials, physical experiments, human adjustment, fundamental limits) with incentives to remove them, past 10% catch-up growth and 30% growth of individual businesses.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Davidson's appendix puts about 25% on such a bottleneck preventing explosive growth. G04/G05 boundary (2026-10-10): kept under G05 because Davidson responds to bottlenecks offered against explosive growth from AI that substitutes for labour; the list also includes regulation and human adjustment, which may act on the pace of automation. The cross-family check of 2026-10-25 read the surrounding objection and response in the 2021-07-02 capture: the objection lists limits that act on automation and limits that remain after full automation (material extraction and transport, physical experiments, fundamental limits), so the response bears on G05, though not on G05 alone.","evidenceIds":["evidence-ec-g05-c"],"attributes":{"thought":{"roles":["earliest-located-tier-2","contesting"],"register":"nonfiction","attested":{"date":"2021-07-02","note":"Earliest retrieved capture."},"engagement":{"kind":"engages-premise","rationale":"Disputes the premise that any one listed bottleneck is certain to bind, while granting \"some chance\" that one does."},"tier":{"value":2,"basis":"Tier 2 (unchanged). Research report by an author with no economics appointment, degree, address or affiliation shown in or near the text and no established record as an economist; the report says it is written to be accessible to non-economists. Under the 2026-10-11 rule the venue does not decide; the author does."}}}},{"id":"ec-m01-sporn-haggerty-1966-comment-b","statement":"Sporn and Haggerty, “Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I” (1966-02) bears on M01. Speaker: Philip Sporn (utility executive; Commission member, non-economist), joined by Patrick E. Haggerty. Sporn says the report under-emphasises technology's positive contributions. A parallel comment by Beirne, Hayes and Reuther (p. 6) says the report lacks urgency.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Sporn's signed general comment, joined by Haggerty; not Commission consensus and not the official report text. Non-economist commission members (tier 2). Matches the cached .txt.","evidenceIds":["evidence-ec-m01-07"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"A comment by two non-economist members, not the Commission's analysis.","tier":{"value":2,"basis":"Comment by two non-economist members (a utility executive and an industrialist)."},"conditions":"Individual members' dissenting-emphasis comment, not Commission text."}}},{"id":"ec-m05-acemoglu-restrepo-2019-wrong-kind-of-ai-b","statement":"Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M05. The authors warn that automation without new tasks would depress labour demand while denying an imminent end of work.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Conditional on no new-task creation; denies near-term end of work.","evidenceIds":["evidence-ec-m05-32"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Considers automation without new tasks, not machines also performing new tasks more cheaply; rejects near-term end of work.","tier":{"value":1,"basis":"NBER working paper (core venue)."}}}},{"id":"lg-oai-andrej-karpathy-founding-member-2015","statement":"OpenAI’s launch announcement of 11 December 2015 names Andrej Karpathy among the group’s founding members.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The predicate means participation in founding; the source wording is “founding members”, a group of research engineers and scientists, not a statement of legal founder status.","evidenceIds":["lg-ev-oai-andrej-karpathy-founding-member-2015"],"attributes":{"role":"Founding member"}},{"id":"lg-goo-shazeer-google-brain-2017","statement":"Attention Is All You Need (arXiv v1, 12 June 2017) lists Noam Shazeer with a Google Brain affiliation.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An author affiliation on a dated paper; it does not establish a title or tenure.","evidenceIds":["lg-ev-goo-shazeer-google-brain-2017"],"attributes":{"role":"author affiliation (Google Brain)"}},{"id":"ap-avi-goldfarb-agrawal-gans-goldfarb-2019-economics-of-ai-introduction","statement":"“Introduction to \"The Economics of Artificial Intelligence: An Agenda\"” (2019-05) credits Avi Goldfarb 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-avi-goldfarb-agrawal-gans-goldfarb-2019-economics-of-ai-introduction"]},{"id":"cmp-faif-2025-d2-definition","statement":"FAIF · December 2025 edition states a provision on the comparison question “How a threshold is defined”, 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. No general definition of what reaching a threshold means; thresholds appear only as benchmark acceptance criteria (“Named levels or thresholds”). The Scope refers to the definition of Catastrophic Risk in California's TFAIA (PDF p. 1, footnote 1) as risk scope, not as a threshold definition. 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-faif-2025-d2-definition-1"],"attributes":{"provision":{"kind":"threshold","key":"comparison-d2-definition","facet":{"id":"D2.definition","status":"partly"}}}},{"id":"fl-intl-seoul-ministerial-thresholds","statement":"The Seoul Ministerial Statement acknowledges the ministers' role, with the private sector, civil society, academia and the international community, in identifying thresholds at which risks from frontier AI models or systems would be severe without appropriate mitigations.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"States a governmental role; the statement sets no threshold.","evidenceIds":["evidence-fl-intl-seoul-ministerial-thresholds-1"],"attributes":{"provision":{"kind":"threshold","key":"intl-seoul-ministerial-thresholds"}}},{"id":"fl-intl-paris-notes-voluntary-commitments","statement":"The Paris statement commends the Bletchley Park and Seoul summits and notes 'the voluntary commitments launched there', without naming any commitment document.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Recorded as what the statement states, not as a link to a commitment text, because the statement names no commitment document.","evidenceIds":["evidence-fl-intl-paris-notes-voluntary-commitments-1"]},{"id":"cpo-pw-google-cfs-offtake","statement":"Google announced on 30 June 2025 an offtake agreement to take 200 MW of carbon-free energy from Commonwealth Fusion Systems' first commercial plant, ARC, to be developed in Chesterfield, Virginia, together with a second capital investment in CFS and an option to buy power from future plants.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"An offtake agreement for a plant not yet built; Google itself states that commercialising fusion is not guaranteed. No delivery date or investment amount is stated in the post. The post does not tie the agreement to a named data centre.","evidenceIds":["evidence-cpo-pw-google-cfs-offtake-1","evidence-cpo-pw-google-cfs-offtake-2","evidence-cpo-pw-google-cfs-offtake-3","evidence-cpo-pw-google-cfs-offtake-4"],"attributes":{"transactionStage":"contracted","quantity":{"value":200,"unit":"megawatts (MW) of carbon free energy","bound":"exact","measures":"contracted_energy"},"power":{"technology":"fusion","mechanism":"ppa"}}},{"id":"ab-arg-d4-sandbrink-articulates","statement":"Sandbrink's preprint (arXiv v1, 24 June 2023) differentiates large language models from biological design tools, stating that language models in particular lower barriers to biological misuse while design tools expand the capabilities of sophisticated actors.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record documents what version 1 of this preprint states; it is not evidence that the proposition is true and implies no influence between texts. Earliest located articulation among the texts searched in this pass; not a claim that no earlier text exists. Urbina and colleagues' 2022 commentary on the dual use of AI-powered drug discovery, an obvious earlier candidate, could not be retrieved (see retrievalFailures) and is not recorded. The abstract's next sentence states what design tools may enable; it is not quoted or paraphrased here, and no agent, method or route is recorded.","evidenceIds":["evidence-ab-arg-d4-sandbrink-articulates-1","evidence-ab-arg-d4-sandbrink-articulates-2","evidence-ab-arg-d4-sandbrink-articulates-3","evidence-ab-arg-d4-sandbrink-articulates-4"],"attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction","qualifier":"Differentiates the two classes by the barriers each affects and by the actors each favours."}}}],"sources":[{"id":"source-openai-launch","title":"Introducing OpenAI","authors":["Greg Brockman","Ilya Sutskever","OpenAI"],"publisher":"OpenAI","publicationDate":"2015-12-11","version":"Dated page, inspected at retrieval","canonicalUrl":"https://openai.com/index/introducing-openai/","evidenceUrl":"https://openai.com/index/introducing-openai/","language":"en","archivedUrl":"https://web.archive.org/web/20260917013950/https://openai.com/index/introducing-openai/","archivedAt":"2026-09-17","accessedAt":"2026-09-17","reuse":{"attribution":"OpenAI, Introducing OpenAI (2015-12-11).","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and minimal attributed excerpts only. Recheck terms before wider redistribution."}},{"id":"source-rsp-3-2","title":"Responsible Scaling Policy — v3.2","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Responsible Scaling Policy v3.2; effective date 2026-04-29.","versionDate":"2026-04-29","canonicalUrl":"https://cdn.sanity.io/files/4zrzovbb/website/28c6241900d90410628a8a2003a5572faae4365a.pdf","evidenceUrl":"https://cdn.sanity.io/files/4zrzovbb/website/28c6241900d90410628a8a2003a5572faae4365a.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260505141510/https://cdn.sanity.io/files/4zrzovbb/website/28c6241900d90410628a8a2003a5572faae4365a.pdf","archivedAt":"2026-05-05","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v3.2.","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-faif-2025","title":"Frontier Artificial Intelligence Framework — December 2025 edition","authors":["xAI"],"publisher":"xAI","publicationDate":null,"version":"Frontier Artificial Intelligence Framework December 2025 edition; revision date 2025-12-30.","versionDate":"2025-12-30","canonicalUrl":"https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","evidenceUrl":"https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260102171656/https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","archivedAt":"2026-01-02","accessedAt":"2026-09-18","reuse":{"attribution":"xAI, Frontier Artificial Intelligence Framework — December 2025 edition.","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-attention-all-you-need-v1","title":"Attention Is All You Need","authors":["Ashish Vaswani","Noam Shazeer","Niki Parmar","Jakob Uszkoreit","Llion Jones","Aidan N. Gomez","Lukasz Kaiser","Illia Polosukhin"],"publisher":"arXiv","publicationDate":"2017-06-12","version":"arXiv:1706.03762v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2017-06-12","canonicalUrl":"https://arxiv.org/abs/1706.03762","evidenceUrl":"https://arxiv.org/pdf/1706.03762v1","language":"en","accessedAt":"2026-09-20","reuse":{"attribution":"Ashish Vaswani; Noam Shazeer; Niki Parmar; Jakob Uszkoreit; Llion Jones; Aidan N. Gomez; Lukasz Kaiser; Illia Polosukhin, Attention Is All You Need (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-summarize-feedback-paper-v1","title":"Learning to summarize from human feedback","authors":["Nisan Stiennon","Long Ouyang","Jeff Wu","Daniel M. Ziegler","Ryan Lowe","Chelsea Voss","Alec Radford","Dario Amodei","Paul Christiano"],"publisher":"arXiv","publicationDate":"2020-09-02","version":"arXiv:2009.01325v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2020-09-02","canonicalUrl":"https://arxiv.org/abs/2009.01325","evidenceUrl":"https://arxiv.org/pdf/2009.01325v1","language":"en","accessedAt":"2026-09-20","reuse":{"attribution":"Nisan Stiennon; Long Ouyang; Jeff Wu; Daniel M. Ziegler; Ryan Lowe; Chelsea Voss; Alec Radford; Dario Amodei; Paul Christiano, Learning to summarize from human feedback (2020-09-02).","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-alibaba-qwen3","title":"Alibaba Introduces Qwen3, Setting New Benchmark in Open-Source AI with Hybrid Reasoning","authors":["Alibaba Group"],"publisher":"Alibaba Group","publicationDate":"2025-04-29","version":"Source language: English (en). Dated group announcement, including Open Access to Drive Innovation section.","canonicalUrl":"https://www.alibabagroup.com/en-US/document-1853940226976645120","evidenceUrl":"https://www.alibabagroup.com/en-US/document-1853940226976645120","language":"en","archivedUrl":"https://web.archive.org/web/20251217011017/https://www.alibabagroup.com/en-US/document-1853940226976645120","archivedAt":"2025-12-17","accessedAt":"2026-09-29","reuse":{"attribution":"Alibaba Group, Alibaba Introduces Qwen3, Setting New Benchmark in Open-Source AI with Hybrid Reasoning (2025-04-29).","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-wh-commitments-2023-09","title":"Voluntary AI Commitments (September 2023)","authors":["The White House"],"publisher":"The White House","publicationDate":"2023-09-12","version":"PDF linked from the fact sheet of 12 September 2023, inspected 2026-10-04 at its archived-site address; byte-identical (SHA-1) to the Internet Archive captures of 2023-09-12 and 2023-09-13 at the original address. The Biden–Harris White House pages are now served from the archived site bidenwhitehouse.archives.gov.","canonicalUrl":"https://bidenwhitehouse.archives.gov/wp-content/uploads/2023/09/Voluntary-AI-Commitments-September-2023.pdf","evidenceUrl":"https://bidenwhitehouse.archives.gov/wp-content/uploads/2023/09/Voluntary-AI-Commitments-September-2023.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20230912193016/https://www.whitehouse.gov/wp-content/uploads/2023/09/Voluntary-AI-Commitments-September-2023.pdf","archivedAt":"2023-09-12","accessedAt":"2026-10-04","reuse":{"attribution":"The White House, Voluntary AI Commitments (September 2023).","licence":"US government work; 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-eu-code-page","title":"The General-Purpose AI Code of Practice (Commission page with the list of signatories)","authors":["European Commission"],"publisher":"European Commission","publicationDate":null,"version":"Policy page inspected live 2026-10-04 (“Last update 31 July 2026”). The list of signatories first appears in the Internet Archive capture of 2025-08-01 (13:04 UTC), with 26 names and the sentence on xAI; the capture of 2025-07-31 (15:27 UTC) has no list. Publisher as printed on the inspected expression: European Commission (Shaping Europe’s digital future).","canonicalUrl":"https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai","evidenceUrl":"https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai","language":"en","archivedUrl":"https://web.archive.org/web/20250801130439/https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai","archivedAt":"2025-08-01","accessedAt":"2026-10-04","reuse":{"attribution":"European Commission, The General-Purpose AI Code of Practice (policy page).","licence":"European Union; 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-eu-code-signing-qa","title":"Signing the General-Purpose AI Code of Practice – Questions & Answers","authors":["European Commission"],"publisher":"European Commission","publicationDate":"2025-07-17","version":"Questions and answers marked “Last update 17 July 2025”, inspected live 2026-10-04; an Internet Archive capture of 2025-07-17 exists and was not compared. Publisher as printed on the inspected expression: European Commission (Shaping Europe’s digital future).","canonicalUrl":"https://digital-strategy.ec.europa.eu/en/faqs/signing-general-purpose-ai-code-practice","evidenceUrl":"https://digital-strategy.ec.europa.eu/en/faqs/signing-general-purpose-ai-code-practice","language":"en","archivedUrl":"https://web.archive.org/web/20250717165024/https://digital-strategy.ec.europa.eu/en/faqs/signing-general-purpose-ai-code-practice","archivedAt":"2025-07-17","accessedAt":"2026-10-04","reuse":{"attribution":"European Commission, Signing the General-Purpose AI Code of Practice – Questions & Answers.","licence":"European Union","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-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-extropians-2000-09-29-yudkowsky-6073","title":"Eugene's nuclear threat (Extropians mailing list)","authors":["Eliezer Yudkowsky"],"publisher":"Extropians mailing list","publicationDate":"2000-09-29","version":"Archive HTML as served, retrieved 2026-10-07 (SHA-256 d8cd64ecadafc77d85742b5123fbb897829fb1f4b1686bb1ad62617c85868e79); quoted lines excluded when reading the author’s text. The served archive shows later URL rewriting and does not authenticate historical bytes. Archive at extropians.weidai.com.","canonicalUrl":"http://extropians.weidai.com/extropians.3Q00/6073.html","evidenceUrl":"http://extropians.weidai.com/extropians.3Q00/6073.html","language":"en","archivedUrl":"https://web.archive.org/web/20261029180714/http://extropians.weidai.com/extropians.3Q00/6073.html","archivedAt":"2026-10-29","accessedAt":"2026-10-07","reuse":{"attribution":"Eliezer Yudkowsky, “Eugene's nuclear threat”, Extropians mailing list, 2000-09-29.","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-lw-2008-fighting-a-rearguard-action-against-the","title":"Fighting a Rearguard Action Against the Truth","authors":["Eliezer Yudkowsky"],"publisher":"Overcoming Bias","publicationDate":"2008-09-24","version":"Post body retrieved through the LessWrong GraphQL API on 2026-10-06 (SHA-256 ca311369027c8858532acd6d0b4da10f803e93a99b2a7c6f5ec9033c76b38bbe), content version 1.0.0. As imported to LessWrong.","canonicalUrl":"https://www.lesswrong.com/posts/PCfaLLtuxes6Jk4S2/fighting-a-rearguard-action-against-the-truth","evidenceUrl":"https://www.lesswrong.com/posts/PCfaLLtuxes6Jk4S2/fighting-a-rearguard-action-against-the-truth","language":"en","archivedUrl":"https://web.archive.org/web/20260414225554/https://www.lesswrong.com/posts/PCfaLLtuxes6Jk4S2/fighting-a-rearguard-action-against-the-truth","archivedAt":"2026-04-14","accessedAt":"2026-10-06","reuse":{"attribution":"Eliezer Yudkowsky, “Fighting a Rearguard Action Against the Truth”, Overcoming Bias, as imported to LessWrong, 2008-09-24.","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-davidson-2021-explosive-economic-growth","title":"Could Advanced AI Drive Explosive Economic Growth?","authors":["Tom Davidson"],"publisher":"Open Philanthropy research report","publicationDate":"2021-06-25","version":"Wayback capture of openphilanthropy.org/could-advanced-ai-drive-explosive-economic-growth (2021-07-02); compared with the 2022 capture of the moved page. Capture retrieved 2026-10-08 (SHA-256 b49aea65d3574a88e914a855aaba57dd4a02728159a187a26a657d0f44ac324b). Published version: Davidson, T. (2021, June 25). Could advanced AI drive explosive economic growth? Open Philanthropy. Report; no other published version recorded.","canonicalUrl":"https://www.openphilanthropy.org/could-advanced-ai-drive-explosive-economic-growth","evidenceUrl":"https://web.archive.org/web/20210702003241/https://www.openphilanthropy.org/could-advanced-ai-drive-explosive-economic-growth","language":"en","archivedUrl":"https://web.archive.org/web/20210702003241/https://www.openphilanthropy.org/could-advanced-ai-drive-explosive-economic-growth","archivedAt":"2021-07-02","accessedAt":"2026-10-08","reuse":{"attribution":"Tom Davidson, “Could Advanced AI Drive Explosive Economic Growth?”, Open Philanthropy research report, 2021-06-25.","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-acemoglu-restrepo-2019-wrong-kind-of-ai","title":"The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand","authors":["Daron Acemoglu","Pascual Restrepo"],"publisher":"NBER Working Paper 25682","publicationDate":"2019-03","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 54ad101804ab7d30ba6d136f413b1c871ffec0a273dbec290c009acd7ad32100.","canonicalUrl":"https://www.nber.org/system/files/working_papers/w25682/w25682.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w25682/w25682.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Daron Acemoglu, Pascual Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand”, NBER Working Paper 25682, 2019-03.","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-agrawal-gans-goldfarb-2019-economics-of-ai-introduction","title":"Introduction to \"The Economics of Artificial Intelligence: An Agenda\"","authors":["Ajay Agrawal","Joshua Gans","Avi Goldfarb"],"publisher":"NBER volume chapter (University of Chicago Press)","publicationDate":"2019-05","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 176dbcbd8a7e924d0f6d7506a9c04fb862a37a99ff3abe07670e7ba892a62839.","canonicalUrl":"https://www.nber.org/system/files/chapters/c14005/c14005.pdf","evidenceUrl":"https://www.nber.org/system/files/chapters/c14005/c14005.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Ajay Agrawal, Joshua Gans, Avi Goldfarb, “Introduction to \"The Economics of Artificial Intelligence: An Agenda\"”, NBER volume chapter (University of Chicago Press), 2019-05.","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-sporn-haggerty-1966-comment","title":"Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I","authors":["Philip Sporn","Patrick E. Haggerty"],"publisher":"Washington, D.C.: U.S. Government Printing Office, February 1966, xiv + 115 pp.; ERIC ED 023 803","publicationDate":"1966-02","version":"Internet Archive OCR text (djvu.txt) of a scanned copy, retrieved 2026-10-09 from https://dn790006.ca.archive.org/0/items/ERICED023803/ERICED023803djvu.txt; OCR errors possible; quotes checked only against clean stretches. Current copy; no earlier capture compared. SHA-256 d23de21a0c05416f957dda13acb680f1baa77ee88e3c094966e8eaba50db637b.","canonicalUrl":"https://archive.org/details/ERIC_ED023803","evidenceUrl":"https://archive.org/details/ERIC_ED023803","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Philip Sporn, Patrick E. Haggerty, “Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I”, Washington, D.C.: U.S. Government Printing Office, February 1966, xiv + 115 pp.; ERIC ED 023 803, 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-seoul-ministerial-statement-2024","title":"Seoul Ministerial Statement for advancing AI safety, innovation and inclusivity: AI Seoul Summit 2024","authors":["Ministers attending the AI Seoul Summit"],"publisher":"Department for Science, Innovation and Technology","publicationDate":"2024-05-22","version":"Retrieved on 2026-10-27 (SHA-256 b719c80a13bdd39fb5f05db9bfd40160833d10741d979202e322a37adc652ac8). GOV.UK HTML policy paper, 'Published 22 May 2024', as served on 2026-10-27. Publisher as printed on the inspected expression: GOV.UK (Department for Science, Innovation and Technology).","canonicalUrl":"https://www.gov.uk/government/publications/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024","evidenceUrl":"https://www.gov.uk/government/publications/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024/seoul-ministerial-statement-for-advancing-ai-safety-innovation-and-inclusivity-ai-seoul-summit-2024","language":"en","accessedAt":"2026-10-27","reuse":{"attribution":"Ministers attending the AI Seoul Summit, “Seoul Ministerial Statement for advancing AI safety, innovation and inclusivity: AI Seoul Summit 2024”, 2024-05-22.","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-paris-ai-action-statement-2025-web","title":"Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet","authors":["Présidence de la République française (Élysée)"],"publisher":"Élysée (Présidence de la République française)","publicationDate":"2025-02-11","version":"Retrieved on 2026-10-27 (SHA-256 6d4bdcba5f263bbe9b3c2b79effda5382df06fef73d849aa5e48bdc1aac15cdf). English web page 'Posted on 11 February 2025', as served on 2026-10-27; the page shows no update date, so when its signatory list took this form is not known Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.elysee.fr/en/emmanuel-macron/2025/02/11/statement-on-inclusive-and-sustainable-artificial-intelligence-for-people-and-the-planet","evidenceUrl":"https://www.elysee.fr/en/emmanuel-macron/2025/02/11/statement-on-inclusive-and-sustainable-artificial-intelligence-for-people-and-the-planet","language":"en","accessedAt":"2026-10-27","reuse":{"attribution":"Présidence de la République française (Élysée), “Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet”, 2025-02-11.","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":"source-cpo-pw-google-cfs-2025","title":"Our latest bet on a fusion-powered future","authors":["Google (The Keyword blog)"],"publisher":"Google (The Keyword blog)","publicationDate":"2025-06-30","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 3dcd96391c17123a0aa2dd7d281a631bb23a178511c0520e09946f83eddedd41). blog.google page as served on 2026-10-28; datePublished 2025-06-30. Retrieved again at admission on 2026-10-30 (SHA-256 85087a4c5ddd5284b88f2fc08a3b9073832dd03ed624360e53a0ed5908bb9f46). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://blog.google/company-news/outreach-and-initiatives/sustainability/our-latest-bet-on-a-fusion-powered-future/","evidenceUrl":"https://blog.google/company-news/outreach-and-initiatives/sustainability/our-latest-bet-on-a-fusion-powered-future/","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Google (The Keyword blog), “Our latest bet on a fusion-powered future”, 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":"source-ab-sandbrink-2023","title":"Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools","authors":["arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford)"],"publisher":"arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford)","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 1ab85d8647f2d23079d35eb8162b70b3f970094106808ef4291fb89615db0afa). arXiv v1 of 24 June 2023 (the article was revised to v8 on 23 December 2023; this record cites v1 only) Retrieved again at admission on 2026-10-30 (SHA-256 1ab85d8647f2d23079d35eb8162b70b3f970094106808ef4291fb89615db0afa, the same bytes).","canonicalUrl":"https://arxiv.org/pdf/2306.13952v1","evidenceUrl":"https://arxiv.org/pdf/2306.13952v1","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford), “Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools”.","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-arxiv-2306-13952-abs","title":"arXiv abstract page for 2306.13952v1 (submission history)","authors":["arXiv (Cornell University)"],"publisher":"arXiv (Cornell University)","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 f05b805a071dd4ad81ec84279296ed530c8a4b679f8d05ffa4760d24a79d8440). Abstract page as served on 2026-10-28; used only for the v1 submission date and the version history Retrieved again at admission on 2026-10-30 (SHA-256 f05b805a071dd4ad81ec84279296ed530c8a4b679f8d05ffa4760d24a79d8440, the same bytes). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://arxiv.org/abs/2306.13952v1","evidenceUrl":"https://arxiv.org/abs/2306.13952v1","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"arXiv (Cornell University), “arXiv abstract page for 2306.13952v1 (submission history)”.","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-lineage-summaries-cites-preferences","sourceRole":"primary","support":"supports","locator":"v1 References [10], PDF p. 11","passage":"Reference [10] identifies Christiano and colleagues’ 2017 paper.","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-cn-alibaba-qwen3-access","sourceRole":"primary","support":"supports","locator":"Open Access to Drive Innovation, first paragraph","passage":"Models are available for download; Model Studio API access will be available soon.","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-absent-rsp-v3-2-no-train-deploy-without-safeguards","sourceRole":"primary","support":"supports","locator":"RSP v3.2 (effective April 29, 2026), whole text layer, PDF pp. 1–20: Introduction (pp. 3–4); §1 with the threshold table (pp. 4–10); §2–§3 Risk Reports (pp. 11–16); §4 Governance (p. 16); Appendix A (pp. 16–17); Appendix B (p. 17); Changelog (pp. 18–20).","passage":"Within the inspected scope, no passage stating the provision “Commitment not to train or deploy without adequate safeguards” was found.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-03","note":"Absence recorded by an automated research assistant on 2026-10-03. Method: Same reading, comparison and case-insensitive search terms as for v3.0, plus a word-level comparison with v3.1 (differences: LTBT-requested external review, LTBT approval of reviewers, LTBT briefings, spacing). No passage stating a commitment not to train or deploy models unless adequate safeguards are implemented, or a commitment to pause training or deployment until such safeguards are in place, was found within the operative text. Hits examined and not counted: the Introduction (p. 3), §1 (p. 4) and Appendix A scenario commitments (p. 17) as in v3.0; the Appendix A introduction sentences on cautionary action and considering pausing (p. 17), as in v3.1; the reproduced RSP v2.0 changelog entry (p. 18). Human editorial review is pending."},{"id":"evidence-prov-wh-2023-09-scope-1","sourceRole":"primary","support":"supports","locator":"Page 1, paragraph beginning “Scope:”","passage":"they apply only to generative models that are overall more powerful than the current most advanced model produced by the company making the commitment","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). The passage was first proposed and machine-verified in the week 4b research pass. Human editorial review is pending."},{"id":"evidence-cm-eu-code-openai-1","sourceRole":"primary","support":"supports","locator":"Section “Signatories of Code of Practice”, in the Internet Archive capture of 2025-08-01","passage":"The list as first published has 26 names, including “OpenAI”.","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-cm-eu-code-openai-2","sourceRole":"primary","support":"context","locator":"Section “Signatories of the code of practice”, live page (last update 31 July 2026)","passage":"The current list of 21 names still includes “OpenAI”.","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-cm-eu-code-openai-3","sourceRole":"primary","support":"context","locator":"Answer to “What does signing the code imply in terms of compliance with the AI Act?”","passage":"The questions and answers say that any opt-out from chapters of the code loses the benefit of facilitating the demonstration of compliance in that respect.","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-ts-a3-moravec-1989-context","sourceRole":"primary","support":"supports","locator":"Opening paragraphs","passage":"Machines will be able to manage their own design and construction.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-06","note":"Paraphrase checked against the retrieved file."},{"id":"evidence-yk-t22-b","sourceRole":"primary","support":"supports","locator":"paragraph beginning 'Eliezer1996-2000's strategies'","passage":"Eliezer1996-2000's strategies had been formed in the total absence of \"Friendly AI\" as a consideration. The whole idea was to get a superintelligence, any superintelligence, as fast as possible","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-06","note":"Candidate pack passage T22-b. The quotation (29 words) matched the retrieved expression (SHA-256 ca311369027c8858532acd6d0b4da10f803e93a99b2a7c6f5ec9033c76b38bbe) when checked on 2026-10-06.","contextOnly":{"reason":"Reports an abandoned belief; never show without the rejection that follows it","requires":"The rejection that follows in the next paragraphs of the same post and in a later post of the same series.","claimIds":["yk-y22-lw-2008-the-magnitude-of-his-own-folly-c"]}},{"id":"evidence-yk-t22-c","passage":"I looked back and I saw the careful arguments I had constructed, for why the wisest choice was to continue forward at full speed, just as I had planned to do before. … Nature was still allowed to say \"So what?\" and kill you.","passageType":"quote"},{"id":"evidence-yk-t24-e2","passageType":"quote","contextOnly":{"reason":"An earlier position later contradicted but not explicitly retracted; never show alone","requires":"His later texts on racing, which do not refer back to this message.","claimIds":["yk-y24-sl4-2007-11-24-yudkowsky-0711-17248-c","yk-y24-time-2023-shut-it-all-down-c"]}},{"id":"evidence-yk-t28-e1","sourceRole":"primary","support":"supports","locator":"Message body, second unquoted paragraph, beginning 'Are you so very, very sure'","passage":"If the people with enough understanding and experience to have created an AI tell you firmly that they have devoted considerable time and effort to Friendliness and they believe the chances are as good as they'll ever get, would you really nuke them?","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-07","note":"Candidate pack passage T28-e1. The quotation (43 words) matched the retrieved expression (SHA-256 d8cd64ecadafc77d85742b5123fbb897829fb1f4b1686bb1ad62617c85868e79) when checked on 2026-10-07.","contextOnly":{"reason":"An earlier view that only partly matches the take; never show alone","requires":"His correction 53 minutes later in the same thread, that good intentions are not enough.","claimIds":["yk-y24-extropians-2000-09-29-yudkowsky-6080-a"]}},{"id":"evidence-ec-g05-c","sourceRole":"primary","support":"supports","locator":"Summary, objection \"Limits to how fast a human economy can grow\", response","passage":"I do think that there is some chance that one of these bottlenecks will prevent explosive growth. On the other hand, no individual bottleneck is certain to apply and there are some reasons to think we could grow at 30% per year","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-08","note":"Candidate pack passage G05-c (research/economics-growth). The quotation (42 words) matched the checked expression (davidson-op-2021) in the pack's verification.","contextOnly":{"reason":"The passage opens with a concession; quoting only the concession or only the rebuttal misstates it.","requires":"Both sentences together."}},{"id":"evidence-ec-m01-07","sourceRole":"primary","support":"supports","locator":"printed p. 7 (general comment on report by Mr. Sporn, joined by Mr. Haggerty)","passage":"Automation represents a logical extension of this 200-year-long history of technological progress … It does not represent a radical departure from past experience","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M01-07 (research/economics-mainstream). The quotation (23 words) matched the retrieved current copy (SHA-256 d23de21a0c05416f957dda13acb680f1baa77ee88e3c094966e8eaba50db637b) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day.","contextOnly":{"reason":"Fresh-context audit (A1) judged that the quotation misleads when shown alone.","requires":"This is a separate general comment by two Commission members (Sporn, joined by Haggerty) objecting to the report's emphasis. It is not the Commission's view."}},{"id":"evidence-ec-m05-32","sourceRole":"primary","support":"supports","locator":"pdf page 7; Varieties of AI (printed p. 5)","passage":"First, if all we do is continue down the path of automation, with no counterbalancing innovations to generate new tasks, the implications for labor are depressing. It will not be the end of work anytime soon","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M05-32 (research/economics-mainstream). The quotation (36 words) matched the retrieved current copy (SHA-256 54ad101804ab7d30ba6d136f413b1c871ffec0a273dbec290c009acd7ad32100) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day.","contextOnly":{"reason":"Fresh-context audit (A3) judged that the quotation misleads when shown alone.","requires":"Conditional scenario ('if all we do is continue down the path of automation...'); the sentence continues with lower labour share and anemic labour demand, which is M06a material."}},{"id":"lg-ev-oai-andrej-karpathy-founding-member-2015","sourceRole":"primary","support":"supports","locator":"Body, paragraph naming the research director, CTO and other founding members","passage":"The announcement lists Andrej Karpathy among “the group’s other founding members”, described as research engineers and scientists.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Andrej Karpathy” 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-goo-shazeer-google-brain-2017","sourceRole":"primary","support":"supports","locator":"Page 1, author and affiliation block","passage":"Noam Shazeer is listed under Google Brain.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Noam Shazeer∗ Google Brain” 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":"evidence-ap-avi-goldfarb-agrawal-gans-goldfarb-2019-economics-of-ai-introduction","sourceRole":"primary","support":"supports","locator":"PDF text layer, first two pages: author name on the title page or first page","passage":"Avi Goldfarb","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":"ev-cmp-faif-2025-d2-definition-1","sourceRole":"primary","support":"supports","locator":"PDF p. 10, 'Operational and Societal Risks' › '5. Deployment Decisions', second paragraph, second sentence","passage":"The edition says that whether, and how far, certain safeguards and mitigations are deployed may depend on a model's performance on relevant benchmarks.","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":"evidence-fl-intl-seoul-ministerial-thresholds-1","sourceRole":"primary","support":"supports","locator":"Safety, paragraph 1.2","passage":"identifying thresholds at which the risks posed by the design, development, deployment and use of frontier AI models or systems would be severe without appropriate mitigations","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (26 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 of the same address (the page’s bytes had changed). Read directly from www.gov.uk. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. The same paragraph recognises a role 'in promoting credible external evaluations' for frontier models developed in their jurisdictions (paraphrase). No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-fl-intl-paris-notes-voluntary-commitments-1","sourceRole":"primary","support":"supports","locator":"Paragraph 5","passage":"We commend the role of the Bletchley Park AI Safety Summit and Seoul Summits that have been essential in progressing international cooperation on AI safety and we note the voluntary commitments launched there.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (33 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 www.elysee.fr. 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 (2026-10-28, item 17) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-google-cfs-offtake-1","sourceRole":"primary","support":"supports","locator":"paragraph beginning “Through our new partnership with CFS…”","passage":"we will offtake 200 megawatts (MW) of carbon free energy from their first commercial plant, to be developed in Chesterfield, Virginia","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (21 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 85087a4c5ddd5284b88f2fc08a3b9073832dd03ed624360e53a0ed5908bb9f46; 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 35) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-google-cfs-offtake-2","sourceRole":"primary","support":"supports","locator":"paragraph beginning “Through our new partnership with CFS…”","passage":"We're also making our second capital investment in CFS","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 85087a4c5ddd5284b88f2fc08a3b9073832dd03ed624360e53a0ed5908bb9f46; 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 35) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-google-cfs-offtake-3","sourceRole":"primary","support":"supports","locator":"paragraph beginning “Building on our initial R&D investment…”","passage":"we have the option to purchase power from future plants","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 (SHA-256 85087a4c5ddd5284b88f2fc08a3b9073832dd03ed624360e53a0ed5908bb9f46; 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 35) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-google-cfs-offtake-4","sourceRole":"primary","support":"context","locator":"paragraph on fusion","passage":"Google writes that commercialising fusion is immensely challenging and success is not guaranteed.","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 35) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-arg-d4-sandbrink-articulates-1","sourceRole":"primary","support":"supports","locator":"Abstract, PDF p. 1","passage":"This article differentiates two classes of AI tools that pose such biosecurity risks: large language models (LLMs) and biological design tools (BDTs).","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (22 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 text is the primary source for its own framing. A cross-family machine check (2026-10-30, item 100) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-arg-d4-sandbrink-articulates-2","sourceRole":"primary","support":"supports","locator":"Abstract, PDF p. 1 (the difference stated)","passage":"Thus, LLMs will in particular lower barriers to biological misuse.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). 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. Paired in the source with the sentence on design tools, quoted as the next item and in the B7 record. A cross-family machine check (2026-10-30, item 100) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-arg-d4-sandbrink-articulates-3","sourceRole":"primary","support":"supports","locator":"Abstract, PDF p. 1 (the design-tool half of the contrast)","passage":"In contrast, BDTs will expand the capabilities of sophisticated actors.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). 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. Added after the audit of 2026-10-28: the statement's second half rested on no cited passage. The same sentence is quoted in recorded separately, so the quotation budget counts it once. A cross-family machine check (2026-10-30, item 100) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-arg-d4-sandbrink-articulates-4","sourceRole":"primary","support":"supports","locator":"Submission history block, v1 entry","passage":"[v1] Sat, 24 Jun 2023","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). 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. Added after the audit of 2026-10-28: the 24 June 2023 date rested on no cited passage. Cited for the submission date only; the retrieved PDF of v1 carries no arXiv stamp. A cross-family machine check (2026-10-30, item 100) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"}],"at":{"entities":[0,1,10,39,47,56,71,92,93,179,186,251,253,270,338,371,376,391,421,508,513,517,531,548,550,582,755,883,1021,1022,1057,1139,1248,1257],"claims":[208,375,487,693,739,819,1137,1138,1150,1174,1276,1299,1382,1574,1584,2244,2676,2953,2955,3176,3416],"sources":[0,41,51,87,93,169,236,244,245,272,358,388,482,500,503,535,853,854,1016,1065,1066],"evidence":[227,398,518,870,938,939,940,1030,1380,1381,1395,1425,1597,1628,1725,1971,1981,2891,3357,3812,3815,4254,4255,4256,4257,4706,4707,4708,4709]},"version":"1lg47t7"}