{"entities":[{"id":"openai","name":"OpenAI"},{"id":"deepmind","name":"DeepMind"},{"id":"eliezer-yudkowsky","name":"Eliezer Yudkowsky"},{"id":"rsp-2023","name":"Responsible Scaling Policy (2023)"},{"id":"fsf-2024","name":"Frontier Safety Framework (2024)"},{"id":"rsp-v2-0","name":"Anthropic Responsible Scaling Policy — v2.0"},{"id":"rsp-v3-3","name":"Anthropic Responsible Scaling Policy — v3.3"},{"id":"preparedness-v2","name":"OpenAI Preparedness Framework — Version 2"},{"id":"thinking-machines-lab","name":"Thinking Machines Lab"},{"id":"attention-all-you-need","name":"Attention Is All You Need"},{"id":"adam-paper","name":"Adam: A Method for Stochastic Optimization"},{"id":"amplification-paper","name":"Supervising strong learners by amplifying weak experts"},{"id":"self-critique-paper","name":"Self-critiquing models for assisting human evaluators"},{"id":"alignment-research-center","name":"Alignment Research Center"},{"id":"gpt4-system-card-2023","name":"GPT-4 System Card"},{"id":"david-silver","name":"David Silver"},{"id":"meta","name":"Meta"},{"id":"meta-faf-v1-1","name":"Meta Frontier AI Framework — Version 1.1"},{"id":"california-sb-53","name":"California SB 53 (2025): Transparency in Frontier Artificial Intelligence Act, as chaptered"},{"id":"inflection-ai","name":"Inflection AI"},{"id":"extropians-1996-09-04-bostrom","name":"Re: Thinking about the future..."},{"id":"fast-or-hard-takeoff","name":"Fast or hard takeoff"},{"id":"human-data-humanlike-mind","name":"Trained on human data, so humanlike"},{"id":"extropians-1999-08-04-yudkowsky-1589","name":"Re: Does a parental model work for having A.I. that can be trusted?"},{"id":"lw-2008-no-universally-compelling-arguments","name":"No Universally Compelling Arguments"},{"id":"ai-as-cheaper-prediction","name":"AI as cheaper prediction"},{"id":"automation-lowers-labour-income","name":"Automation lowers the labour share, wages or employment"},{"id":"compensating-policy-for-ai","name":"Compensating policy for AI’s distributional effects"},{"id":"acemoglu-restrepo-2018-ai-automation-and-work","name":"Artificial Intelligence, Automation and Work"},{"id":"agrawal-gans-goldfarb-2019-ambiguous-labor-market-impact","name":"Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction"},{"id":"leontief-duchin-1984-impacts-of-automation","name":"The Impacts of Automation on Employment, 1963–2000. Final Report"},{"id":"machine-intelligence-exceeds-humans-in-all-domains","name":"Machine intelligence exceeding human cognition in essentially all domains"},{"id":"hanson-2011-debating-yudkowsky","name":"Debating Yudkowsky"},{"id":"andrew-tulloch","name":"Andrew Tulloch"},{"id":"bytedance-seed","name":"ByteDance Seed / 字节跳动 Seed"},{"id":"meta-superintelligence-labs","name":"Meta Superintelligence Labs"},{"id":"mia-glaese","name":"Mia Glaese"},{"id":"reid-hoffman","name":"Reid Hoffman"},{"id":"seed-20-model-card","name":"Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity"},{"id":"seed-ap-announcement-2026","name":"Seed Alignment & Preparedness 全球顶尖人才招募启动"},{"id":"us-eo-14299-2025","name":"Executive Order 14299: Deploying Advanced Nuclear Reactor Technologies for National Security"},{"id":"stargate-port-washington","name":"Lighthouse campus, Port Washington, Wisconsin (Stargate)"},{"id":"vantage-data-centers","name":"Vantage Data Centers"},{"id":"openai-chatgpt-agent-system-card-2025","name":"ChatGPT Agent System Card"}],"claims":[{"id":"lineage-attention-uses-adam","statement":"Attention Is All You Need reports using Adam to train its models.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Author-reported use in the specified study, not independent replication, general institutional adoption, sole origin or proof of effectiveness. The target identifies a work; its inspected version is not necessarily the version used by these authors.","evidenceIds":["evidence-lineage-attention-uses-adam"],"attributes":{"method":"Adam optimiser with the paper’s specified learning-rate schedule"}},{"id":"lineage-ii-critique-cites-amplification","statement":"Self-critiquing models for assisting human evaluators cites Supervising strong learners by amplifying weak experts.","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 citation alone does not show that the experiments implement the full proposal.","evidenceIds":["evidence-lineage-ii-critique-cites-amplification"]},{"id":"eval-i-gpt4-arc-credit","statement":"OpenAI’s GPT-4 system card credits ARC with preliminary autonomous-replication evaluations.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The source names ARC. It says ARC lacked fine-tuning and final deployed-model access; results do not certify GPT-4-launch safety.","evidenceIds":["evidence-eval-i-gpt4-arc-credit"]},{"id":"atlas-david-silver-deepmind","statement":"The AlphaFold paper lists David Silver with a DeepMind affiliation.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Paper affiliation at publication, not an employment interval or current appointment.","evidenceIds":["evidence-atlas-david-silver-deepmind"],"attributes":{"role":"Paper affiliation"}},{"id":"prov-meta-faf-v1-1-critical-threshold","statement":"Frontier AI Framework v1.1 states that a model is at the critical risk threshold when it would uniquely enable the execution of at least one threat scenario identified as potentially sufficient to produce a catastrophic outcome and that risk cannot be mitigated in the proposed deployment context.","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. Prose elsewhere in Version 1.1 words the critical threshold slightly differently (§3.2 adds that the risk cannot be further mitigated; §3.3 omits the mitigation clause).","evidenceIds":["evidence-prov-meta-faf-v1-1-critical-threshold"],"attributes":{"provision":{"kind":"threshold","key":"meta-critical-threshold"}}},{"id":"prov-meta-faf-v1-1-high-deployment-measure","statement":"Frontier AI Framework v1.1 states that if evaluation results indicate a frontier AI has a “high” risk threshold by providing significant uplift towards realization of a catastrophic outcome, Meta will not release it externally.","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. Quoted from the 2025-02-13 expression; the revised expression captured from 2025-03-28 reads “towards realization of a threat scenario” at the same point.","evidenceIds":["evidence-prov-meta-faf-v1-1-high-deployment-measure"],"attributes":{"provision":{"kind":"required_action","key":"meta-high-deployment-measure"}}},{"id":"provchange-rsp-v2-0-policy-change-approval","statement":"v2.0 states that policy changes are proposed by the CEO and the Responsible Scaling Officer and approved by the Board of Directors in consultation with the LTBT; v1.0's Update Process named approval by the board of directors following consultation with the LTBT and named no proposer.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: RSP · v2.0 states the provision as quoted in its provision record (PDF p. 16 (printed p. 12), §7.1 Internal Governance, numbered item 7 ('Policy changes'), first sentence); RSP · 2023 states it as quoted in its provision record (PDF p. 10 (printed p. 10), 'Procedural Commitments', numbered item 1, first sentence). The difference described is read directly from the two passages.","editorialState":"seed","qualification":"A textual difference between the two editions’ wording of this keyed provision. It is not an assessment of whether the change strengthened or weakened the framework, and not evidence of implementation.","evidenceIds":["evidence-provchange-rsp-v2-0-policy-change-approval-later","evidence-provchange-rsp-v2-0-policy-change-approval-earlier"],"attributes":{"changeBasis":"document_comparison","provision":{"kind":"decision_authority","key":"rsp-policy-change-approval"}}},{"id":"sb-framework-modification-30-days","statement":"SB 53 requires a large frontier developer that materially modifies its frontier AI framework to clearly and conspicuously publish the modified framework, with a justification for the modification, within 30 days.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the enacted text states, quoted from the chaptered expression. It is not evidence that the statute applies to any named developer, or of compliance, enforcement or effect. Neither § 22757.12(b)(2) nor the chapter's definitions (§ 22757.11) define 'material modification'. The phrase appears elsewhere only in § 22757.11(i)(2), about computing applied to models. The subdivision does not expressly state the event from which the 30 days are counted. Whether a particular framework edit was material, and how the period is counted, are matters of legal interpretation that this record does not resolve. This record is not evidence that any developer published a modified framework on time.","evidenceIds":["evidence-sb-framework-modification-30-days-1","evidence-sb-framework-modification-30-days-2"],"attributes":{"provision":{"kind":"required_action","key":"sb53-framework-modification-publication"}}},{"id":"ts-a4-bostrom-1996-09-04-context","statement":"Bostrom’s message of 1996-09-04 says access to a nanotechnology laboratory might be all an AI needs to take off, the contest being over before anyone but the AI had realised it had begun.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A rapid scenario without the criterion’s period.","evidenceIds":["evidence-ts-a4-bostrom-1996-09-04-context"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"A rapid \"take off\" ends a contest before anyone notices, but no period and no \"hard\" or \"fast\" takeoff is named."}}},{"id":"yk-y21-extropians-1999-08-04-yudkowsky-1589-c","statement":"The Extropians message “Re: Does a parental model work for having A.I. that can be trusted?” (1999-08-04) contests Y21. In a thread titled 'Does a parental model work for having A.I. that can be trusted?', he tells the poster that how AIs are raised makes no difference and that 'parenting' cannot help, because it depends on human emotions an AI will not have.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Premise addressed: Training on human text or behaviour yields human-like thought and motives. The premise addressed is upbringing or parenting of an AI, not training on human text corpora (cf. his SL4 message of 2002-03-09). He posted the reply twice within a minute (also as a second Extropians message, ending 'that won't exist in an AI'). Quotation shortened to respect the ~10% cap for a short message. Near-duplicate of that second Extropians message, posted 15 seconds apart.","evidenceIds":["evidence-yk-t21-e1"],"attributes":{"thought":{"roles":["contesting","earliest-located"],"register":"nonfiction","engagement":{"kind":"engages-premise","rationale":"Under the message-level rule, the passage, read with the paragraph, message or dialogue turn it belongs to, gives a reason against the premise that training on human text or behaviour yields human-like thought and motives."}}}},{"id":"ec-m04-agrawal-gans-goldfarb-2019-ambiguous-labor-market-impact-a","statement":"Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) articulates M04. The authors identify current AI with prediction and draw labour-market consequences from cheaper prediction for complementary decision tasks. The authors set prediction as a perfect complement to decisions, so falling prediction costs change returns to labour and capital in decision tasks.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"(First passage) NBER working paper (February 2019) of the JEP 2019 article. (Second passage) Modelling assumption of the paper.","evidenceIds":["evidence-ec-m04-08","evidence-ec-m04-09"],"attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","tier":{"value":1,"basis":"NBER working paper (core venue)."},"conditions":"(M04-08) AI equated with machine-learning prediction; decisions and judgement as complements. (M04-09) Prediction and decision tasks as perfect complements."}}},{"id":"ec-m06a-acemoglu-restrepo-2018-ai-automation-and-work-a","statement":"Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) articulates M06a. Even with offsetting forces, automation lowers the labour share.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"NBER Working Paper 24196, January 2018 (NBER conference paper; later in The Economics of Artificial Intelligence). \"in- creases\" is a line-break artefact in the text layer (the PDF reads \"increases\").","evidenceIds":["evidence-ec-m06a-13"],"attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","tier":{"value":1,"basis":"NBER working paper (core venue)."},"conditions":"Countervailing productivity, capital-accumulation and deepening effects present; absent new labour-intensive tasks."}}},{"id":"ec-m06b-leontief-duchin-1984-impacts-of-automation-b","statement":"Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M06b. Speaker: Wassily Leontief and Faye Duchin (economists). The authors call their study 'a first systematic albeit partial glance' at employment prospects nearly twenty years ahead.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Matches the cached .txt. Policy remarks elsewhere concern education and training (computer-based courseware).","evidenceIds":["evidence-ec-m06b-08"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Names income distribution as a next research step (citing Duchin 1984); proposes no redistribution or compensating policy.","tier":{"value":1,"basis":"Report by economists (Leontief and Duchin, NYU Institute for Economic Analysis)."},"conditions":"Part of a list of 'next steps in our agenda', with a closed dynamic model determining consumption and employment together."}}},{"id":"ec-e02-hanson-2011-debating-yudkowsky-a","statement":"Hanson, “Debating Yudkowsky” (2011-07-03), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E02. Hanson's summary position after the 2011 debate.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Hanson's own statement. He affirms superintelligences \"with vast and broad mental capacities\" but gives no explicit comparison with humans across all domains. Tier 2 blog post.","evidenceIds":["evidence-ec-e02-03"],"attributes":{"thought":{"roles":["earliest-located-tier-2"],"register":"nonfiction","tier":{"value":2,"basis":"Blog debate (2008 Overcoming Bias posts) between an economist and a non-economist: tier 2 as an online debate; the maintainer confirmed tier 2 for the debate on 2026-10-12."}}}},{"id":"lg-deal-hoffman-cofounded-inflection-2022","statement":"Mustafa Suleyman announced on 8 March 2022 that he and Reid were co-founding Inflection AI; Inflection AI's press release of 21 July 2026 names Reid Hoffman as a Co-Founder of Inflection AI.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The 2022 post names only 'Reid' (a Greylock general partner); the full name rests on Inflection's July 2026 press release, which names Reid Hoffman as Co-Founder. Neither source states an operating role at Inflection or an incorporation date.","evidenceIds":["lg-ev-deal-hoffman-cofounded-inflection-2022","lg-ev-deal-hoffman-cofounder-inflection-2026"],"attributes":{"role":"Co-Founder"}},{"id":"lg-mx-msl-unit-of-meta-2025","statement":"Mark Zuckerberg's memo of 30 June 2025, as published by CNBC, names Meta's overall AI organisation Meta Superintelligence Labs (MSL).","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting: the memo text as published by CNBC; Meta's own copy was not retrieved.","evidenceIds":["lg-ev-mx-msl-unit-of-meta-2025"]},{"id":"lg-new-tulloch-tml-departure-2025","statement":"Reuters reported on 11 October 2025, relaying a Wall Street Journal report, that Andrew Tulloch had left Thinking Machines Lab to join Meta Platforms.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on second-hand reporting: Reuters relaying the Wall Street Journal, retrieved as a TradingView syndication; the Wall Street Journal article and any company statement were not retrieved. No reason is recorded; no Meta title is established.","evidenceIds":["lg-ev-new-tulloch-tml-departure-2025","lg-ev-new-tulloch-tml-departure-2025-cofounder"],"attributes":{"role":"co-founder"}},{"id":"lg-gs-glaese-openai-head-of-alignment-2026","statement":"WIRED reported on 10 July 2026 that, according to a memo by OpenAI's chief research officer Mark Chen, OpenAI's safety teams would report to Mia Glaese, whom WIRED identifies as the company's VP of research and head of alignment.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting (WIRED, relaying an internal memo it had seen). It does not establish when Glaese took up either title. The title wording may be WIRED's rather than the memo's.","evidenceIds":["lg-ev-gs-glaese-openai-head-of-alignment-2026"],"attributes":{"role":"VP of research and head of alignment"}},{"id":"ap-eliezer-yudkowsky-lw-2008-no-universally-compelling-arguments","statement":"“No Universally Compelling Arguments” (2008-06-26) credits Eliezer Yudkowsky as its author.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Credited authorship of this expression; it says nothing about priority for the ideas discussed.","evidenceIds":["evidence-ap-eliezer-yudkowsky-lw-2008-no-universally-compelling-arguments"]},{"id":"cmp-rsp-v3-3-d8-changes","statement":"RSP · v3.3 states a provision on the comparison question “Stated changes from the previous edition”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Observed in the text: the novel chemical/biological row is redefined; §3.1 adds “Off-cycle updates” and narrows the internal models in scope to those that could pose risks of high-stakes misalignment or automated R&D significantly above those of models in a prior Risk Report; and the sabotage row is retitled “Misaligned AI systems in high-stakes settings”. What the edition states, paraphrased from the cited expression with its page. It is not evidence of implementation, compliance or effectiveness, and a shared question does not make two editions’ provisions equivalent.","evidenceIds":["ev-cmp-rsp-v3-3-d8-changes-1"],"attributes":{"provision":{"kind":"revision","key":"comparison-d8-changes","facet":{"id":"D8.changes","status":"stated"}}}},{"id":"cmp-fsf-2024-d7-internal","statement":"No passage meeting the comparison question “Internal reporting” was found in FSF · 2024 within the inspected scope.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: the comparison question asks for the following. Reporting or records the edition requires inside the developer (to leadership, a board, staff or a compliance function). The inspected scope of FSF · 2024 was: Whole text layer, PDF pp. 1–7 (all sections, both tables, footnotes and acknowledgements read). Method: the edition was read in full and its text layer searched case-insensitively for report, record, log, inform, notif/alert, review, council. Hits examined and not counted: PDF p. 3 “enhanced logging” and the PDF p. 4 measure of automated monitoring and logging for internal deployments at scale are security or deployment mitigations, not reporting to leadership or a compliance function. The Future work passage on exploring policies for alerting stakeholder bodies (PDF p. 6) concerns external bodies (considered under “Public or external reporting”). “inform” hits (PDF pp. 1, 6) concern the Framework being informed by external discussion, or forecasting to inform preparations. No “report” or “record” hits. The conclusion is bounded by that scope and method.","editorialState":"seed","qualification":"A recorded absence within the stated scope: not a claim that the provision was deleted, and not a statement about practice or intent.","evidenceIds":["ev-cmp-fsf-2024-d7-internal"],"attributes":{"provision":{"kind":"reporting","key":"comparison-d7-internal","facet":{"id":"D7.internal"}}}},{"id":"c2-bytedance-seed-20-model-card","statement":"ByteDance Seed posted the Seed2.0 model card to arXiv on 30 June 2026 (version 1).","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The card's comparisons with other developers' models are not recorded. Its contributors are listed alphabetically and, as the card states, some names are internal aliases, so no individual authorship is drawn from the list.","evidenceIds":["c2-ev-bytedance-seed-20-model-card","c2-ev-bytedance-seed-20-model-card-stamp","c2-ev-bytedance-seed-20-model-card-credit","c2-ev-bytedance-seed-20-model-card-aliases"]},{"id":"c2-bytedance-seed-ap-announcement","statement":"ByteDance Seed's announcement of 14 October 2026 lists among the planned work of its Alignment & Preparedness department drafting model behaviour specifications and safety guidelines, building a safety evaluation system, and promoting third-party professional testing.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Planned work stated in a recruitment announcement; nothing here establishes that any specification, evaluation system or third-party test exists or has been carried out. 制定 is rendered 'draft/formulate' and 推动 'promote'.","evidenceIds":["c2-ev-bytedance-seed-ap-announcement","c2-ev-bytedance-seed-ap-announcement-third-party"]},{"id":"cpo-pw-us-eo14299-ai-dc-defense","statement":"Section 4(a) of Executive Order 14299 of 23 May 2025 directed the Secretary of Energy to initiate the process for designating AI data centres within the 48 contiguous States and the District of Columbia that are located at or operated in coordination with DOE facilities as critical defense facilities, “where appropriate”, and stated that the electrical infrastructure needed to operate them shall be considered defense critical electric infrastructure, “as appropriate and consistent with applicable law”.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A direction to begin a process, with three limits the section states: the geographic limit (the 48 contiguous States and the District of Columbia), “where appropriate” for the designation, and “as appropriate and consistent with applicable law” for the treatment of the electrical infrastructure. No designation is recorded here (cross-family check, item 49).","evidenceIds":["evidence-cpo-pw-us-eo14299-ai-dc-defense-1","evidence-cpo-pw-us-eo14299-ai-dc-defense-2","evidence-cpo-pw-us-eo14299-ai-dc-defense-3","evidence-cpo-pw-us-eo14299-ai-dc-defense-4"],"attributes":{"provision":{"kind":"required_action","key":"us-eo14299-sec4a"}}},{"id":"cpo-dc-vantage-port-washington-spend","statement":"Vantage's release described its own \"$15B+ investment\" in Port Washington.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Vantage's stated investment, not the project's total cost and not OpenAI's or Oracle's spending. A plan, not spending incurred.","evidenceIds":["evidence-cpo-dc-vantage-port-washington-spend-1"],"attributes":{"transactionType":"infrastructure_spend","relationshipContext":"external","transactionStage":"planned","amount":{"value":15000000000,"currency":"USD","bound":"more_than","scope":"project","scopeNote":"Vantage's investment, as stated"}}},{"id":"ab-fw-pf-2-bio-chem-category-definition","statement":"The Preparedness Framework Version 2 describes its Biological and Chemical tracked category as capabilities that, in addition to unlocking discoveries and cures, can also reduce barriers to creating and using biological or chemical weapons.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record states how the edition describes a tracked category. It is not a finding about any model.","evidenceIds":["evidence-ab-fw-pf-2-bio-chem-category-definition-1"],"attributes":{"provision":{"kind":"risk_domain","key":"pf-bio-chem-category-definition"}}},{"id":"ab-ev-chatgpt-agent-high-capability","statement":"The ChatGPT Agent System Card states that OpenAI decided to treat the launch as High capability in the Biological and Chemical domain under its Preparedness Framework, activating the associated safeguards, while stating that it does not have definitive evidence that the model could meaningfully help a novice to create severe biological harm and has taken a precautionary approach.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A developer's own classification decision under its own framework, stated as precautionary; it is neither an independent finding that the threshold was crossed nor evidence that the safeguards activated are effective. The threshold is recorded at the generality of the card's own wording; no scenario, step or agent is recorded.","evidenceIds":["evidence-ab-ev-chatgpt-agent-high-capability-1","evidence-ab-ev-chatgpt-agent-high-capability-2","evidence-ab-ev-chatgpt-agent-high-capability-3"],"attributes":{"evaluatedSystem":"ChatGPT agent, as released on 17 July 2025","evaluation":{"what":"Whether the model reaches the High capability threshold in the Biological and Chemical domain of OpenAI's Preparedness Framework","how":"A capabilities assessment across the card's biological and chemical evaluations, reviewed under the framework by OpenAI's Safety Advisory Group","result":"Treated as High capability in the Biological and Chemical domain, with the associated safeguards activated, as a precautionary approach; the card states the decision is informed by an incremental improvement relative to previous models","caveats":"The card states there is no definitive evidence that the model could meaningfully help a novice to create severe biological harm, which is its defined High capability threshold, and that its evaluations represent a lower bound for potential capabilities."}}}],"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-2-0","title":"Responsible Scaling Policy — v2.0","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Responsible Scaling Policy v2.0; effective date 2024-10-15. v2.0 identity is supplied by the official archive and later dated changelogs; the cover is not numbered.","versionDate":"2024-10-15","canonicalUrl":"https://www-cdn.anthropic.com/616dee633636e5bd309cb73aed8622e80fe47839.pdf","evidenceUrl":"https://www-cdn.anthropic.com/616dee633636e5bd309cb73aed8622e80fe47839.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20241022154907/https://assets.anthropic.com/m/24a47b00f10301cd/original/Anthropic-Responsible-Scaling-Policy-2024-10-15.pdf","archivedAt":"2024-10-22","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v2.0.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-rsp-3-3","title":"Responsible Scaling Policy — v3.3","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Responsible Scaling Policy v3.3; effective date 2026-05-26.","versionDate":"2026-05-26","canonicalUrl":"https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","evidenceUrl":"https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260528170023/https://cdn.sanity.io/files/4zrzovbb/website/c11e84981d0a7281a1b229f3fa6af0da66eaf43f.pdf","archivedAt":"2026-05-28","accessedAt":"2026-09-18","reuse":{"attribution":"Anthropic, Responsible Scaling Policy — v3.3.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-fsf-1-0","title":"Frontier Safety Framework — v1.0","authors":["Google DeepMind"],"publisher":"Google DeepMind","publicationDate":"2024-05-17","version":"Frontier Safety Framework v1.0; publication date 2024-05-17.","versionDate":"2024-05-17","canonicalUrl":"https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/introducing-the-frontier-safety-framework/fsf-technical-report.pdf","evidenceUrl":"https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/introducing-the-frontier-safety-framework/fsf-technical-report.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20240517140902/https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/introducing-the-frontier-safety-framework/fsf-technical-report.pdf","archivedAt":"2024-05-17","accessedAt":"2026-09-18","reuse":{"attribution":"Google DeepMind, Frontier Safety Framework — 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-pf-2","title":"Preparedness Framework — Version 2","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":"2025-04-15","version":"Preparedness Framework Version 2; revision date 2025-04-15.","versionDate":"2025-04-15","canonicalUrl":"https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf","evidenceUrl":"https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20250415191354/https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf","archivedAt":"2025-04-15","accessedAt":"2026-09-18","reuse":{"attribution":"OpenAI, Preparedness Framework — Version 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-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-self-critique-paper-v1","title":"Self-critiquing models for assisting human evaluators","authors":["William Saunders","Catherine Yeh","Jeff Wu","Steven Bills","Long Ouyang","Jonathan Ward","Jan Leike"],"publisher":"arXiv","publicationDate":"2022-06-12","version":"arXiv:2206.05802v1; first-submission date from version history. Later revisions are not substituted.","versionDate":"2022-06-12","canonicalUrl":"https://arxiv.org/abs/2206.05802","evidenceUrl":"https://arxiv.org/pdf/2206.05802v1","language":"en","accessedAt":"2026-09-25","reuse":{"attribution":"William Saunders; Catherine Yeh; Jeff Wu; Steven Bills; Long Ouyang; Jonathan Ward; Jan Leike, Self-critiquing models for assisting human evaluators.","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-eval-i-gpt4-card","title":"GPT-4 System Card","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":null,"version":"60-page PDF expression retrieved 2026-09-27; printed pagination starts at 41. No publication/revision date established from this PDF. Re-inspected: 60-page PDF expression reinspected 2026-09-28; printed pagination starts at 41. Publication/revision date unknown. (Merged on 2026-10-29 with the record of the re-inspection, source-eval-ontology-i-gpt4-card, by consistency audit fix B4.)","canonicalUrl":"https://cdn.openai.com/papers/gpt-4-system-card.pdf","evidenceUrl":"https://cdn.openai.com/papers/gpt-4-system-card.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260927044417/https://cdn.openai.com/papers/gpt-4-system-card.pdf","archivedAt":"2026-09-27","accessedAt":"2026-09-28","reuse":{"attribution":"OpenAI, GPT-4 System Card.","licence":"No open redistribution licence asserted","note":"Attributed paraphrases and bibliographic metadata only; original documents linked, not redistributed."}},{"id":"source-atlas-alphafold-paper","title":"Highly accurate protein structure prediction with AlphaFold","authors":["John Jumper","Richard Evans","Alexander Pritzel","Tim Green","Michael Figurnov","Olaf Ronneberger","Kathryn Tunyasuvunakool","Russ Bates","Augustin Žídek","Anna Potapenko","Alex Bridgland","Clemens Meyer","Simon A. A. Kohl","Andrew J. Ballard","Andrew Cowie","Bernardino Romera-Paredes","Stanislav Nikolov","Rishub Jain","Jonas Adler","Trevor Back","Stig Petersen","David Reiman","Ellen Clancy","Michal Zielinski","Martin Steinegger","Michalina Pacholska","Tamas Berghammer","Sebastian Bodenstein","David Silver","Oriol Vinyals","Andrew W. Senior","Koray Kavukcuoglu","Pushmeet Kohli","Demis Hassabis"],"publisher":"Nature","publicationDate":"2021-07-15","version":"Nature 596, 583–589 (2021), DOI 10.1038/s41586-021-03819-2; version of record HTML.","canonicalUrl":"https://www.nature.com/articles/s41586-021-03819-2","evidenceUrl":"https://www.nature.com/articles/s41586-021-03819-2","language":"en","archivedUrl":"https://web.archive.org/web/20260923202227/https://www.nature.com/articles/s41586-021-03819-2","archivedAt":"2026-09-23","accessedAt":"2026-09-28","reuse":{"attribution":"John Jumper; Richard Evans; Alexander Pritzel; et al.. Highly accurate protein structure prediction with AlphaFold.","licence":"No redistribution licence asserted.","note":"Short attributed paraphrases only; original publication is linked, not redistributed."}},{"id":"source-meta-faf-v1-1","title":"Frontier AI Framework — Version 1.1","authors":["Meta"],"publisher":"Meta","publicationDate":"2025-02-03","version":"Frontier AI Framework, Version 1.1, as captured by the Internet Archive on 2025-02-13, the earliest complete capture located. The original address has returned HTTP 500 since April 2026; a revised expression under the same label was captured from 2025-03-28 and is recorded separately.","versionDate":"2025-02-03","canonicalUrl":"https://ai.meta.com/static-resource/meta-frontier-ai-framework/","evidenceUrl":"https://web.archive.org/web/20250213105620/https://ai.meta.com/static-resource/meta-frontier-ai-framework/","language":"en","archivedUrl":"https://web.archive.org/web/20250213105620/https://ai.meta.com/static-resource/meta-frontier-ai-framework/","archivedAt":"2025-02-13","accessedAt":"2026-10-03","reuse":{"attribution":"Meta, Frontier AI Framework, Version 1.1 (2025).","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-openai-chatgpt-agent-system-card","title":"ChatGPT Agent System Card","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":"2025-07-17","version":"System card dated 17 July 2025 on its cover. The quoted sentence and its page match the PDF served on 2026-10-04 and the Internet Archive capture of 2025-07-17; the two copies differ only in chart labels.","canonicalUrl":"https://openai.com/index/chatgpt-agent-system-card/","evidenceUrl":"https://cdn.openai.com/pdf/839e66fc-602c-48bf-81d3-b21eacc3459d/chatgpt_agent_system_card.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20250717225133/https://cdn.openai.com/pdf/839e66fc-602c-48bf-81d3-b21eacc3459d/chatgpt_agent_system_card.pdf","archivedAt":"2025-07-17","accessedAt":"2026-10-04","reuse":{"attribution":"OpenAI, ChatGPT Agent System Card (2025-07-17).","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-ca-sb53-chaptered","title":"SB-53 Artificial intelligence models: large developers. (2025-2026) — bill text, chaptered version (Chapter 138, Statutes of 2025)","authors":["California Legislature"],"publisher":"California State Legislature","publicationDate":"2025-09-29","version":"Chaptered version dated 09/29/25, inspected live 2026-10-04; quote also matched in the 2025-09-30 capture. Publisher as printed on the inspected expression: California Legislature (California Legislative Information, Office of Legislative Counsel).","canonicalUrl":"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53","evidenceUrl":"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53","language":"en","archivedUrl":"https://web.archive.org/web/20250930154105/https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53","archivedAt":"2025-09-30","accessedAt":"2026-10-04","reuse":{"attribution":"California Legislature, SB-53 Artificial intelligence models: large developers. (2025-2026) — bill text, chaptered version (Chapter 138, Statutes of 2025).","licence":"Official government text; no redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-ca-bpc-22757-12","title":"Business and Professions Code § 22757.12 — codified text (California Legislative Information)","authors":["California Legislature"],"publisher":"California State Legislature","publicationDate":null,"version":"Codified section page inspected live 2026-10-04 (history note present, no amendment note). CDX lists the 2026-01-30 capture with status 200, but its content could not be re-retrieved during adjudication (HTTP 429). A verifier reported that the capture shows the same history note. Publisher as printed on the inspected expression: California Legislature (California Legislative Information, Office of Legislative Counsel).","canonicalUrl":"https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=BPC&sectionNum=22757.12.","evidenceUrl":"https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=BPC&sectionNum=22757.12.","language":"en","archivedUrl":"https://web.archive.org/web/20260130073345/https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=BPC&sectionNum=22757.12","archivedAt":"2026-01-30","accessedAt":"2026-10-04","reuse":{"attribution":"California Legislature, Business and Professions Code § 22757.12 — codified text (California Legislative Information).","licence":"Official government text; no redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-extropians-1996-09-04-bostrom","title":"Re: Thinking about the future... (Extropians mailing list)","authors":["Nick Bostrom"],"publisher":"Extropians mailing list","publicationDate":"1996-09-04","version":"Archive HTML as served, retrieved 2026-10-06 with the hash recorded by the 1996–2022 candidate search; quoted lines excluded when reading the author’s text. Archive at extropians.weidai.com. Date: the archive’s receipt line reads “Wed Sep  4 10:09:44 1996 MST” and the message is filed in the archive’s 1996 volume, which give the record’s date; the sent header is garbled and ends “04 Sep 95”, which nothing else in the message supports (receipt line re-read in a direct retrieval on 2026-10-27).","canonicalUrl":"http://extropians.weidai.com/extropians.96/0764.html","evidenceUrl":"http://extropians.weidai.com/extropians.96/0764.html","language":"en","archivedUrl":"https://web.archive.org/web/20150427180540/http://extropians.weidai.com/extropians.96/0764.html","archivedAt":"2015-04-27","accessedAt":"2026-10-06","reuse":{"attribution":"Nick Bostrom, \"Re: Thinking about the future...\", Extropians mailing list, 1996-09-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-extropians-1999-08-04-yudkowsky-1589","title":"Re: Does a parental model work for having A.I. that can be trusted? (Extropians mailing list)","authors":["Eliezer Yudkowsky"],"publisher":"Extropians mailing list","publicationDate":"1999-08-04","version":"Archive HTML as served, retrieved 2026-10-07 (SHA-256 6b82e1b19b4c1a8d6a6ec54b78b4c6b0d662f4c23f05cf46b61462e4c3e0335d); 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.3Q99/1589.html","evidenceUrl":"http://extropians.weidai.com/extropians.3Q99/1589.html","language":"en","archivedUrl":"https://web.archive.org/web/20250514232432/https://extropians.weidai.com/extropians.3Q99/1589.html","archivedAt":"2025-05-14","accessedAt":"2026-10-07","reuse":{"attribution":"Eliezer Yudkowsky, “Re: Does a parental model work for having A.I. that can be trusted?”, Extropians mailing list, 1999-08-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-lw-2008-no-universally-compelling-arguments","title":"No Universally Compelling Arguments","authors":["Eliezer Yudkowsky"],"publisher":"Overcoming Bias","publicationDate":"2008-06-26","version":"Post body retrieved through the LessWrong GraphQL API on 2026-10-06 (SHA-256 437193a5a1540d060d8bd3c1d10d89831afae39e30d17637357f108f93246232), content version 1.1.0; quotations also compared with the posting-day version 1.0.0. As imported to LessWrong.","canonicalUrl":"https://www.lesswrong.com/posts/PtoQdG7E8MxYJrigu/no-universally-compelling-arguments","evidenceUrl":"https://www.lesswrong.com/posts/PtoQdG7E8MxYJrigu/no-universally-compelling-arguments","language":"en","archivedUrl":"https://web.archive.org/web/20260920015821/https://www.lesswrong.com/posts/PtoQdG7E8MxYJrigu/no-universally-compelling-arguments","archivedAt":"2026-09-20","accessedAt":"2026-10-06","reuse":{"attribution":"Eliezer Yudkowsky, “No Universally Compelling Arguments”, Overcoming Bias, as imported to LessWrong, 2008-06-26.","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-2018-ai-automation-and-work","title":"Artificial Intelligence, Automation and Work","authors":["Daron Acemoglu","Pascual Restrepo"],"publisher":"NBER Working Paper 24196","publicationDate":"2018-01","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 0707870c32273416f9e92849a4eb539c74a6febbbd318e94e3d7eacab24eaea5.","canonicalUrl":"https://www.nber.org/system/files/working_papers/w24196/w24196.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w24196/w24196.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Daron Acemoglu, Pascual Restrepo, “Artificial Intelligence, Automation and Work”, NBER Working Paper 24196, 2018-01.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-agrawal-gans-goldfarb-2019-ambiguous-labor-market-impact","title":"Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction","authors":["Ajay Agrawal","Joshua S. Gans","Avi Goldfarb"],"publisher":"NBER Working Paper 25619","publicationDate":"2019-02","version":"Full text. Current copy retrieved directly on 2026-10-09; no archive capture (Wayback unreachable), so earlier bytes are not authenticated. SHA-256 dbd015967a47f77a41ff8028f5cc8208c3bb0e8cb44d6b1cc9be04f7df9c2450. Journal version not retrieved. Agrawal, A., Gans, J. S., & Goldfarb, A. (2019). Artificial intelligence: The ambiguous labor market impact of automating prediction. Journal of Economic Perspectives, 33(2), 31–50. https://doi.org/10.1257/jep.33.2.31","canonicalUrl":"https://www.nber.org/system/files/working_papers/w25619/w25619.pdf","evidenceUrl":"https://www.nber.org/system/files/working_papers/w25619/w25619.pdf","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Ajay Agrawal, Joshua S. Gans, Avi Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction”, NBER Working Paper 25619, 2019-02.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-leontief-duchin-1984-impacts-of-automation","title":"The Impacts of Automation on Employment, 1963–2000. Final Report","authors":["Wassily Leontief","Faye Duchin"],"publisher":"New York: Institute for Economic Analysis, New York University; National Science Foundation contract PRA-8012844; April 1984, 376 pp.; ERIC ED 241 743","publicationDate":"1984-04","version":"Internet Archive OCR text (djvu.txt) of a scanned copy, retrieved 2026-10-09 from https://dn711508.ca.archive.org/0/items/microIA411534480031/microIA411534480031djvu.txt; OCR errors possible; quotes checked only against clean stretches. Current copy; no earlier capture compared. SHA-256 cd03df677ca1a723f843eccec0fa1415b3227d604bc83c6783a40321b9a785e9. Book version not retrieved; the 1984 report was. Leontief, W., & Duchin, F. (1986). The future impact of automation on workers. Oxford University Press.","canonicalUrl":"https://archive.org/details/micro_IA41153448_0031","evidenceUrl":"https://archive.org/details/micro_IA41153448_0031","language":"en","accessedAt":"2026-10-09","reuse":{"attribution":"Wassily Leontief, Faye Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report”, New York: Institute for Economic Analysis, New York University; National Science Foundation contract PRA-8012844; April 1984, 376 pp.; ERIC ED 241 743, 1984-04.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-hanson-2011-debating-yudkowsky-miri-2013","title":"Debating Yudkowsky","authors":["Robin Hanson"],"publisher":"Machine Intelligence Research Institute, The Hanson-Yudkowsky AI-Foom Debate (2013)","publicationDate":"2013","version":"Chapter 59 of The Hanson-Yudkowsky AI-Foom Debate (Machine Intelligence Research Institute, 2013), which compiles the 2008 Overcoming Bias debate with later posts; the compilation heads the post with its date. Current MIRI PDF, retrieved 2026-10-11. SHA-256 6b6dfd0416402e1776d7ddcece95298df2df0ed9bc0bd4a6f4bd81c4b1e5102a.","canonicalUrl":"https://intelligence.org/files/AIFoomDebate.pdf","evidenceUrl":"https://intelligence.org/files/AIFoomDebate.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260929131419/https://intelligence.org/files/AIFoomDebate.pdf","archivedAt":"2026-09-29","accessedAt":"2026-10-11","reuse":{"attribution":"Robin Hanson, “Debating Yudkowsky”, in Machine Intelligence Research Institute, The Hanson-Yudkowsky AI-Foom Debate (2013).","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-greylock-inflection-2022","title":"A New Paradigm in Human-Machine Interaction","authors":["Mustafa Suleyman"],"publisher":"Greylock","publicationDate":"2022-03-08","version":"Dated post, retrieved on 2026-10-15 (SHA-256 015830d0e5a70a73bbea8bd3b986e1ddc92e467624208bfbbffc93bb1d5eafb3) directly from the publisher.","canonicalUrl":"https://greylock.com/blog/a-new-paradigm-in-human-machine-interaction/","evidenceUrl":"https://greylock.com/blog/a-new-paradigm-in-human-machine-interaction/","language":"en","archivedUrl":"https://web.archive.org/web/20261003053005/https://greylock.com/blog/a-new-paradigm-in-human-machine-interaction/","archivedAt":"2026-10-03","accessedAt":"2026-10-15","reuse":{"attribution":"Mustafa Suleyman, “A New Paradigm in Human-Machine Interaction”, 2022-03-08.","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":"lg-src-deal-inflection-labs-2026","title":"Inflection AI is Shaping the Future of Personal Intelligence","authors":["Inflection AI"],"publisher":"Inflection AI","publicationDate":"2026-07-21","version":"Dated page, retrieved on 2026-10-15 (SHA-256 964d63cb84d49f505f2017319f480b943988e86e271b6f159c4cde427b703857) directly from the publisher.","canonicalUrl":"https://inflection.ai/blog/inflection-ai-is-shaping-the-future-of-personal-intelligence","evidenceUrl":"https://inflection.ai/blog/inflection-ai-is-shaping-the-future-of-personal-intelligence","language":"en","archivedUrl":"https://web.archive.org/web/20260903012551/https://inflection.ai/blog/inflection-ai-is-shaping-the-future-of-personal-intelligence","archivedAt":"2026-09-03","accessedAt":"2026-10-15","reuse":{"attribution":"Inflection AI, “Inflection AI is Shaping the Future of Personal Intelligence”, 2026-07-21.","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-seed-ap-2026","title":"Seed Alignment & Preparedness 全球顶尖人才招募启动","authors":["ByteDance Seed"],"publisher":"ByteDance Seed (字节跳动 Seed)","publicationDate":"2026-10-14","version":"Dated page, retrieved on 2026-10-15 (SHA-256 7f8c724c943ffd9d8b23bf1f6943b792b29088e4b5478c11ae903f8c38eee031) directly from the publisher.","canonicalUrl":"https://seed.bytedance.com/zh/blog/seed-alignment-preparedness-team-looking-for-global-talents","evidenceUrl":"https://seed.bytedance.com/zh/blog/seed-alignment-preparedness-team-looking-for-global-talents","language":"zh","archivedUrl":"https://web.archive.org/web/20261015054822/https://seed.bytedance.com/zh/blog/seed-alignment-preparedness-team-looking-for-global-talents","archivedAt":"2026-10-15","accessedAt":"2026-10-15","reuse":{"attribution":"ByteDance Seed, “Seed Alignment & Preparedness 全球顶尖人才招募启动”, 2026-10-14.","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-gs-wired-heidecke-2026","title":"OpenAI’s Head of Safety Is Leaving the Company","authors":["Maxwell Zeff"],"publisher":"WIRED","publicationDate":"2026-07-10","version":"Dated page, retrieved on 2026-10-15 (SHA-256 09db3f87023f04a986538909b8318cb8ee6ff4871a3fcc339edabb3ec6c02b2a) directly from the publisher.","canonicalUrl":"https://www.wired.com/story/openai-head-of-safety-leaving/","evidenceUrl":"https://www.wired.com/story/openai-head-of-safety-leaving/","language":"en","archivedUrl":"https://web.archive.org/web/20260929154344/https://www.wired.com/story/openai-head-of-safety-leaving/","archivedAt":"2026-09-29","accessedAt":"2026-10-15","reuse":{"attribution":"Maxwell Zeff, “OpenAI’s Head of Safety Is Leaving the Company”, 2026-07-10.","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":"lg-src-mx-cnbc-msl-memo","title":"Mark Zuckerberg announces creation of Meta Superintelligence Labs. Read the memo","authors":["CNBC"],"publisher":"CNBC","publicationDate":"2025-06-30","version":"Dated page, retrieved on 2026-10-15 (SHA-256 3c940666760a1f8148ddb1674b0cd4c5f8d28a5eaa32f79fc7052c8cce1ca6b4) through the third-party reader relay r.jina.ai, which renders the publisher’s current page (the publisher refused direct retrieval); not an archive copy.","canonicalUrl":"https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","evidenceUrl":"https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","language":"en","archivedUrl":"https://web.archive.org/web/20261015225529/https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html","archivedAt":"2026-10-15","accessedAt":"2026-10-15","reuse":{"attribution":"CNBC, “Mark Zuckerberg announces creation of Meta Superintelligence Labs. Read the memo”, 2025-06-30.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-new-reuters-tulloch-2025","title":"Thinking Machines Lab co-founder Tulloch departs for Meta, WSJ reports (Reuters wire as syndicated by TradingView)","authors":["Reuters (via TradingView)"],"publisher":"Reuters (via TradingView)","publicationDate":"2025-10-11","version":"Dated page, retrieved on 2026-10-15 (SHA-256 64d8bc7c8e4515119a62db123f8f671b937383db12034429ec10d3efcf2fa93f) directly from the publisher.","canonicalUrl":"https://vn.tradingview.com/news/reuters.com%2C2025%3Anewsml_L2N3VS042%3A0-thinking-machines-lab-co-founder-tulloch-departs-for-meta-wsj-reports/","evidenceUrl":"https://vn.tradingview.com/news/reuters.com%2C2025%3Anewsml_L2N3VS042%3A0-thinking-machines-lab-co-founder-tulloch-departs-for-meta-wsj-reports/","language":"en","archivedUrl":"https://web.archive.org/web/20251012020713/https://www.tradingview.com/news/reuters.com,2025:newsml_L2N3VS042:0-thinking-machines-lab-co-founder-tulloch-departs-for-meta-wsj-reports/","archivedAt":"2025-10-12","accessedAt":"2026-10-15","reuse":{"attribution":"Reuters (via TradingView), “Thinking Machines Lab co-founder Tulloch departs for Meta, WSJ reports (Reuters wire as syndicated by TradingView)”, 2025-10-11.","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":"c2-src-seed-20-arxiv-v1","title":"Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity","authors":["Bytedance Seed"],"publisher":"arXiv","publicationDate":"2026-06-30","version":"Dated document, retrieved on 2026-10-27 (SHA-256 1449a46c4ebcdf4c6e00c863d15cdf6418ee75cfa8d1de91453cbf62d1f4a441) directly from the publisher. Version: 2607.00248v1. Version 1 as submitted on 30 June 2026.","canonicalUrl":"https://arxiv.org/pdf/2607.00248v1","evidenceUrl":"https://arxiv.org/pdf/2607.00248v1","language":"en","accessedAt":"2026-10-27","reuse":{"attribution":"Bytedance Seed, “Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity”, 2026-06-30.","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-cpo-pw-eo-14299-fr","title":"Executive Order 14299: Deploying Advanced Nuclear Reactor Technologies for National Security (90 FR 22581)","authors":["The White House"],"publisher":"Federal Register (Office of the Federal Register, National Archives) via govinfo.gov","publicationDate":"2025-05-29","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 22ddba09a2e933ddc8cf0c82c11641358f620da7566bc693aac46e6a6434a4b7). Federal Register print of 29 May 2025, Vol. 90, No. 102, pp. 22581 ff., FR Doc. 2025-09796. The PDF text layer scatters the title words. Retrieved again at admission on 2026-10-30 (SHA-256 22ddba09a2e933ddc8cf0c82c11641358f620da7566bc693aac46e6a6434a4b7, the same bytes).","canonicalUrl":"https://www.govinfo.gov/content/pkg/FR-2025-05-29/pdf/2025-09796.pdf","evidenceUrl":"https://www.govinfo.gov/content/pkg/FR-2025-05-29/pdf/2025-09796.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"The White House, “Executive Order 14299: Deploying Advanced Nuclear Reactor Technologies for National Security (90 FR 22581)”, 2025-05-29.","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-cpo-dc-vantage-stargate-wisconsin-2025","title":"OpenAI, Oracle and Vantage Data Centers Announce Stargate Data Center Site in Wisconsin","authors":["Vantage Data Centers"],"publisher":"Vantage Data Centers","publicationDate":"2025-10-22","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 4649d2b2ed48a67388661a3ff494dbe40ccc52ec3e250001932b8b1bf83b2027). News page as served on 2026-10-28; dateline \"DENVER (Oct. 22, 2025)\". The DigitalBridge investor-relations PDF copy refused (HTTP 403). Retrieved again at admission on 2026-10-30 (SHA-256 0aea6b3ef87fcfa11f074347e414fcc95f3f6900c910440b8829929f9cb0fd50). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://vantage-dc.com/news/openai-oracle-and-vantage-data-centers-announce-stargate-data-center-site-in-wisconsin/","evidenceUrl":"https://vantage-dc.com/news/openai-oracle-and-vantage-data-centers-announce-stargate-data-center-site-in-wisconsin/","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Vantage Data Centers, “OpenAI, Oracle and Vantage Data Centers Announce Stargate Data Center Site in Wisconsin”, 2025-10-22.","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."}}],"evidence":[{"id":"evidence-lineage-attention-uses-adam","sourceRole":"primary","support":"supports","locator":"v1 §5.3, PDF p. 7","passage":"The training section names Adam as the optimiser used.","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-critique-cites-amplification","sourceRole":"primary","support":"supports","locator":"PDF §1.1, p. 1; p. 21, [CSA18]","passage":"The introduction cites this proposal as an antecedent of AI-assisted human evaluation.","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 citation alone does not show that the experiments implement the full proposal."},{"id":"evidence-eval-i-gpt4-arc-credit","sourceRole":"primary","support":"supports","locator":"Section 2.9, printed pp55–56 (PDF pages 15–16), paragraphs on ARC access and preliminary experiments.","passage":"OpenAI reports ARC tested early GPT-4 versions and cautions that the deployed version still required evaluation.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-09-27","note":"Original publisher passage machine-inspected. This is not human review or independent reproduction of the reported activity."},{"id":"evidence-atlas-david-silver-deepmind","sourceRole":"primary","support":"supports","locator":"Authors and Affiliations, entry 1 (DeepMind, London, UK)","passage":"David Silver appears under the DeepMind affiliation.","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-prov-meta-faf-v1-1-critical-threshold","sourceRole":"primary","support":"supports","locator":"PDF p. 13 (printed p. 13), §3.3 Risk thresholds, Table 1: Risk thresholds for frontier AI, Critical row, Risk Threshold column (text below the measure label 'Stop development')","passage":"The model would uniquely enable the execution of at least one of the threat scenarios that have been identified as potentially sufficient to produce a catastrophic outcome and that risk cannot be mitigated in the proposed deployment context.","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 2025-02-13, the earliest complete capture (the file is no longer served). Extraction note: Verbatim in both copies (meta-faf-a, 2025-02-13 capture; meta-faf-b, later variant). Prose wording varies: §3.3 (PDF p. 12) assigns the critical threshold where the model 'would uniquely enable execution of a threat scenario', with no mitigation clause; §3.2 (PDF p. 11) assigns it if the model does so 'and cannot be further mitigated'. Human editorial review is pending."},{"id":"evidence-prov-meta-faf-v1-1-high-deployment-measure","sourceRole":"primary","support":"supports","locator":"PDF p. 17 (printed p. 17), §4.2 Evaluation and mitigation, paragraph beginning 'We may take into account monetary costs', second sentence","passage":"If the results of our evaluations indicate that a frontier AI has a “high” risk threshold by providing significant uplift towards realization of a catastrophic outcome we will not release the frontier AI externally.","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 2025-02-13, the earliest complete capture (the file is no longer served). The quote is not verbatim in meta-faf-b.txt; see the extraction note. Extraction note: Cited expression is the 2025-02-13 copy (meta-faf-a). Not verbatim in meta-faf-b: the later variant reads 'towards realization of a threat scenario' where meta-faf-a reads 'towards realization of a catastrophic outcome'; the rest of the sentence is identical. Table 1 (PDF p. 13) labels the High measure 'Do not release'. Human editorial review is pending."},{"id":"evidence-provchange-rsp-v2-0-policy-change-approval-later","sourceRole":"primary","support":"supports","locator":"PDF p. 16 (printed p. 12), §7.1 Internal Governance, numbered item 7 ('Policy changes'), first sentence","passage":"RSP v2.0 states that changes to the policy will be proposed by the CEO and the Responsible Scaling Officer and approved by the Board of Directors, in consultation with the Long-Term Benefit Trust.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-03","note":"Paraphrase checked by an automated research assistant on 2026-10-03 against the cited edition; the edition’s own wording is quoted in its provision record. Human editorial review is pending."},{"id":"evidence-provchange-rsp-v2-0-policy-change-approval-earlier","sourceRole":"primary","support":"supports","locator":"PDF p. 10 (printed p. 10), 'Procedural Commitments', numbered item 1, first sentence","passage":"RSP v1.0 states that Anthropic will follow an Update Process for the policy that includes approval by the board of directors following consultation with the Long-Term Benefit Trust.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-03","note":"Paraphrase checked by an automated research assistant on 2026-10-03 against the cited edition; the edition’s own wording is quoted in its provision record. Human editorial review is pending."},{"id":"evidence-sb-framework-modification-30-days-1","sourceRole":"primary","support":"supports","locator":"SEC. 2, Bus. & Prof. Code § 22757.12(b)(2)","passage":"publish the modified frontier AI framework and a justification for that modification within 30 days.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim excerpt proposed by a research agent and checked by two independent fresh-context verifier agents and decided by a third adjudicating agent on 2026-10-04; re-matched mechanically by the integrator against a direct fetch of the cited expression. The statement paraphrases the condition, a material modification by a large frontier developer. Human editorial review is pending."},{"id":"evidence-sb-framework-modification-30-days-2","sourceRole":"primary","support":"context","locator":"History note following § 22757.12","passage":"(Added by Stats. 2025, Ch. 138, Sec. 2. (SB 53) Effective January 1, 2026.)","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-04","note":"Verbatim excerpt proposed by a research agent and checked by two independent fresh-context verifier agents and decided by a third adjudicating agent on 2026-10-04; re-matched mechanically by the integrator against a direct fetch of the cited expression. Legislative Counsel's annotation on the official codes site. This item supports only the effective date. Context: the codified section and its history note as displayed when checked; the provision itself is quoted from the chaptered text. Human editorial review is pending."},{"id":"evidence-ts-a4-bostrom-1996-09-04-context","sourceRole":"primary","support":"supports","locator":"Message body, first unquoted paragraph of the reply","passage":"Access to a nanotechnology laboratory might be all the AI would need \"in order to take off\"; the contest would be over before anyone but the AI noticed it had begun.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-06","note":"Paraphrase, so that the message’s quotation limit covers the A3 and A6 passages; locator corrected after a machine check."},{"id":"evidence-yk-t21-e1","sourceRole":"primary","support":"supports","locator":"message body, paragraph beginning 'You're being far too anthropomorphic.…'","passage":"a base of human emotions that isn't going to exist in an AI.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-07","note":"Candidate pack passage T21-e1. The quotation (13 words) matched the retrieved expression (SHA-256 6b82e1b19b4c1a8d6a6ec54b78b4c6b0d662f4c23f05cf46b61462e4c3e0335d) when checked on 2026-10-07."},{"id":"evidence-ec-m04-08","sourceRole":"primary","support":"supports","locator":"pdf page 3; introduction (printed p. 1)","passage":"Machine learning does not represent an increase in artificial general intelligence of the kind that could substitute machines for all aspects of human cognition, but rather one particular aspect of intelligence: prediction","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M04-08 (research/economics-mainstream). The quotation (32 words) matched the retrieved current copy (SHA-256 dbd015967a47f77a41ff8028f5cc8208c3bb0e8cb44d6b1cc9be04f7df9c2450) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day."},{"id":"evidence-ec-m04-09","sourceRole":"primary","support":"supports","locator":"pdf page 3; introduction (printed p. 1)","passage":"Prediction has no value in the absence of a decision. In this sense, each prediction task is a perfect complement to a decision task.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M04-09 (research/economics-mainstream). The quotation (24 words) matched the retrieved current copy (SHA-256 dbd015967a47f77a41ff8028f5cc8208c3bb0e8cb44d6b1cc9be04f7df9c2450) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day."},{"id":"evidence-ec-m06a-13","sourceRole":"primary","support":"supports","locator":"pdf page 2; abstract","passage":"Even when they are strong, automation in- creases output per worker more than wages and reduce the share of labor in national income.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M06a-13 (research/economics-mainstream). The quotation (23 words) matched the retrieved current copy (SHA-256 0707870c32273416f9e92849a4eb539c74a6febbbd318e94e3d7eacab24eaea5) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day."},{"id":"evidence-ec-m06b-08","sourceRole":"primary","support":"supports","locator":"printed p. 1.35","passage":"investigate the impacts on the distribution of income implied by the technological assumptions","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-14","note":"Candidate pack passage M06b-08 (research/economics-mainstream). The quotation (13 words) matched the retrieved current copy (SHA-256 cd03df677ca1a723f843eccec0fa1415b3227d604bc83c6783a40321b9a785e9) in the pack's verification. Re-checked at admission on 2026-10-14 against a copy retrieved that day.","contextOnly":{"reason":"Fresh-context audit (A2) judged that the quotation misleads when shown alone.","requires":"The text lists future research steps; it does not make a policy call. The same page notes that computer-based education will become an item of government policy."}},{"id":"evidence-ec-e02-03","sourceRole":"primary","support":"supports","locator":"\"Debating Yudkowsky\", 3 July 2011, MIRI compilation (2013) ch. 59, post body; pdf p. 509","passage":"Future superintelligences will exist, but their vast and broad mental capacities will come mainly from vast mental content and computational resources.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-16","note":"Candidate pack passage E02-03 (research/economics-mainstream/esoteric). The quotation (21 words) matched the retrieved current copy (SHA-256 6b6dfd0416402e1776d7ddcece95298df2df0ed9bc0bd4a6f4bd81c4b1e5102a) in the pack's verification of 2026-10-15. Re-checked at admission on 2026-10-16 against a copy retrieved that day.","contextOnly":{"reason":"Fresh-context audit (A1) judged that the quotation misleads when shown alone.","requires":"In context this sentence leads into Hanson's argument against architecture-driven foom ('By comparison, their general architectural innovations will be minor additions'). 'Vast and broad mental capacities' implies exceeding humans but does not say 'all domains'."}},{"id":"lg-ev-deal-hoffman-cofounded-inflection-2022","sourceRole":"primary","support":"supports","locator":"Paragraph 3","passage":"Reid and I are excited to announce that we are co-founding a new company, Inflection AI.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (16 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. 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":"direct"},{"id":"lg-ev-deal-hoffman-cofounder-inflection-2026","sourceRole":"primary","support":"supports","locator":"Quotation attributed to Reid Hoffman","passage":"said Reid Hoffman, Co-Founder of Inflection AI","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (7 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. 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":"direct"},{"id":"lg-ev-mx-msl-unit-of-meta-2025","sourceRole":"secondary","support":"supports","locator":"Memo, paragraph 2","passage":"We're going to call our overall organization Meta Superintelligence Labs (MSL).","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-new-tulloch-tml-departure-2025","sourceRole":"secondary","support":"supports","locator":"Paragraph 1","passage":"Reuters reported, attributing it to the Wall Street Journal, that Tulloch had left the company to join Meta Platforms.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “left the company to join Meta Platforms” 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. 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":"direct"},{"id":"lg-ev-new-tulloch-tml-departure-2025-cofounder","sourceRole":"secondary","support":"context","locator":"Paragraph 1","passage":"Reuters described Andrew Tulloch as the co-founder of Mira Murati’s Thinking Machines Lab.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Andrew Tulloch, the co-founder” 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. 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":"direct"},{"id":"lg-ev-gs-glaese-openai-head-of-alignment-2026","sourceRole":"secondary","support":"supports","locator":"Paragraph 2","passage":"WIRED writes that, in a memo to staff it had seen, Mark Chen said OpenAI's safety teams would report to the company's VP of research and head of alignment, Mia Glaese.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “VP of research and head of alignment Mia Glaese” 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. 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":"direct"},{"id":"evidence-ap-eliezer-yudkowsky-lw-2008-no-universally-compelling-arguments","sourceRole":"primary","support":"supports","locator":"Post author list in the publisher’s post data (LessWrong GraphQL API)","passage":"Eliezer Yudkowsky","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 through the publisher’s post data and checked by a reviewer from a different model family on 2026-10-27. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"ev-cmp-rsp-v3-3-d8-changes-1","sourceRole":"primary","support":"supports","locator":"PDF p. 20 (printed p. 19), Changelog, entry 'May 26, 2026 (RSP v3.3)'","passage":"The v3.3 changelog entry lists three changes: a revised novel chemical/biological weapons threshold, intended to follow the threat model of concern more closely; a refined approach to off-cycle updates on individual models’ risks; and minor terminology changes.","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-fsf-2024-d7-internal","sourceRole":"primary","support":"supports","locator":"FSF · 2024: Whole text layer, PDF pp. 1–7 (all sections, both tables, footnotes and acknowledgements read)","passage":"Within the inspected scope, no passage meeting the comparison question “Internal reporting” was found.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-26","note":"Absence recorded by an automated research assistant on 2026-10-15 for the framework comparison, audited in a fresh context, checked by a reviewer from a different model family on 2026-10-15 and re-checked on 2026-10-26. Human editorial review is pending."},{"id":"c2-ev-bytedance-seed-20-model-card","sourceRole":"primary","support":"supports","locator":"Title page, title","passage":"Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (9 words), a part of a longer passage that the pack joined across separate fields of the page, matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The pack had joined this field with others of the same page into one quotation; at admission each field became its own quotation (cross-family check, 2026-10-27). No human editorial review.","retrievalRoute":"direct"},{"id":"c2-ev-bytedance-seed-20-model-card-stamp","sourceRole":"primary","support":"supports","locator":"Page 1, arXiv stamp","passage":"arXiv:2607.00248v1 [cs.AI] 30 Jun 2026","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (5 words), a part of a longer passage that the pack joined across separate fields of the page, matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The pack had joined this field with others of the same page into one quotation; at admission each field became its own quotation (cross-family check, 2026-10-27). No human editorial review.","retrievalRoute":"direct"},{"id":"c2-ev-bytedance-seed-20-model-card-credit","sourceRole":"primary","support":"supports","locator":"Title page, author credit","passage":"Bytedance Seed","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (2 words), a part of a longer passage that the pack joined across separate fields of the page, matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The pack had joined this field with others of the same page into one quotation; at admission each field became its own quotation (cross-family check, 2026-10-27). No human editorial review.","retrievalRoute":"direct"},{"id":"c2-ev-bytedance-seed-20-model-card-aliases","sourceRole":"primary","support":"context","locator":"Section 7, Contributors","passage":"The authors are listed in alphabetical order by their first names. Some names refer to the authors' internal aliases at the company.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (22 words) matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"c2-ev-bytedance-seed-ap-announcement","sourceRole":"primary","support":"supports","locator":"List of work areas, item 'Improve the safety foundation and workflows'","passage":"制定模型行为规范和安全准则，建立科学的安全评估体系","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (24 words, each CJK character counted as one) matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. English machine translation, not verified by a person: “Formulate model behaviour specifications and safety guidelines, and establish a scientific safety evaluation system”. No human editorial review.","retrievalRoute":"direct"},{"id":"c2-ev-bytedance-seed-ap-announcement-third-party","sourceRole":"primary","support":"supports","locator":"List of work areas, item 'Verify and improve model safety with external partners'","passage":"推动第三方专业测试，与高校和研究机构共同推进前沿安全能力研究","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Candidate pack passage, Chinese frontier-AI institutions, second collection (research/china-labs-2). The quotation (29 words, each CJK character counted as one) matched the retrieved text on 2026-10-27 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. English machine translation, not verified by a person: “Promote third-party professional testing, and advance frontier safety-capability research together with universities and research institutions”. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-eo14299-ai-dc-defense-1","sourceRole":"primary","support":"supports","locator":"Sec. 4(a), 90 FR 22582","passage":"initiate the process for designating AI data centers within the 48 contiguous States and the District of Columbia, in whole or in part, that are located at or operated in coordination with Department of Energy facilities","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (36 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 49) re-read the record’s sources and found partial support; its corrections were applied at admission. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-eo14299-ai-dc-defense-2","sourceRole":"primary","support":"supports","locator":"Sec. 4(a), 90 FR 22582","passage":"shall be considered defense critical electric infrastructure","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (7 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval 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 49) re-read the record’s sources and found partial support; its corrections were applied at admission. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-eo14299-ai-dc-defense-3","sourceRole":"primary","support":"supports","locator":"Sec. 4(a), 90 FR 22582","passage":"as critical defense facilities, where appropriate","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Added at the admission of the compute, power and orbit pack (research/compute-power-orbit). The limit of the designation that the section states (cross-family check, item 49). The quotation (6 words) matched a direct retrieval of 2026-10-30 (SHA-256 22ddba09a2e933ddc8cf0c82c11641358f620da7566bc693aac46e6a6434a4b7). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-eo14299-ai-dc-defense-4","sourceRole":"primary","support":"supports","locator":"Sec. 4(a), 90 FR 22582","passage":"as appropriate and consistent with applicable law","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Added at the admission of the compute, power and orbit pack (research/compute-power-orbit). The applicable-law limit that the section states for the treatment of the electrical infrastructure (cross-family check, item 49). The quotation (7 words) matched a direct retrieval of 2026-10-30 (SHA-256 22ddba09a2e933ddc8cf0c82c11641358f620da7566bc693aac46e6a6434a4b7). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-dc-vantage-port-washington-spend-1","sourceRole":"primary","support":"supports","locator":"subheading","passage":"Vantage’s $15B+ investment in Port Washington","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (6 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 0aea6b3ef87fcfa11f074347e414fcc95f3f6900c910440b8829929f9cb0fd50; 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 25) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-fw-pf-2-bio-chem-category-definition-1","sourceRole":"primary","support":"supports","locator":"PDF p. 2 (printed p. 1), opening page, first bullet of the Tracked Categories list","passage":"Biological and Chemical capabilities that, in addition to unlocking discoveries and cures, can also reduce barriers to creating and using biological or chemical weapons.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (24 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-chatgpt-agent-high-capability-1","sourceRole":"primary","support":"supports","locator":"PDF p. 5 (printed p. 4), §1 Introduction, final paragraph","passage":"We have decided to treat this launch as High capability in the Biological and Chemical domain under our Preparedness Framework, activating the associated safeguards.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (24 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-30, item 34) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-chatgpt-agent-high-capability-2","sourceRole":"primary","support":"context","locator":"PDF p. 5 (printed p. 4), §1 Introduction, final paragraph","passage":"While we do not have definitive evidence that this model could meaningfully help a novice to create severe biological harm - our defined threshold for High capability - we have chosen to take a precautionary approach.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (34 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. The card's own qualification of the decision. A cross-family machine check (2026-10-30, item 34) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-chatgpt-agent-high-capability-3","sourceRole":"primary","support":"context","locator":"PDF p. 16 (printed p. 15), §5.1 Capabilities Assessment, introductory paragraph","passage":"evaluations represent a lower bound for potential capabilities","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (8 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A caveat the card gives for all of its capability evaluations. A cross-family machine check (2026-10-30, item 34) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"}],"at":{"entities":[0,9,19,21,22,42,48,54,82,87,89,99,102,108,113,131,157,199,205,235,291,339,370,402,423,516,518,519,541,547,579,586,594,629,643,741,742,774,992,993,1074,1177,1198,1244],"claims":[202,231,260,301,465,468,592,608,872,1135,1353,1401,1423,1504,1619,1776,1805,1881,2439,2576,2597,2838,2847,3049,3159,3267,3389],"sources":[35,36,42,44,49,87,101,114,134,178,190,191,194,292,369,390,493,499,532,542,596,598,619,668,700,728,817,934,984],"evidence":[221,252,281,322,496,499,663,664,701,702,1086,1377,1690,1691,1747,1770,1877,2040,2041,2284,2322,2323,2430,3086,3234,3261,3629,3630,3631,3632,3646,3647,3972,3973,3974,3975,4206,4481,4655,4656,4657]},"version":"1lg47t7"}