{"entities":[{"id":"anthropic","name":"Anthropic"},{"id":"deepmind","name":"DeepMind"},{"id":"eliezer-yudkowsky","name":"Eliezer Yudkowsky"},{"id":"google-research","name":"Google Research"},{"id":"rsp-v2-0","name":"Anthropic Responsible Scaling Policy — v2.0"},{"id":"faif-2025","name":"xAI Frontier Artificial Intelligence Framework — December 2025 edition"},{"id":"alibaba-group","name":"Alibaba Group / 阿里巴巴集团"},{"id":"jingren-zhou","name":"Jingren Zhou"},{"id":"ms-fgf-2025","name":"Microsoft Frontier Governance Framework — Version 1 (February 2025)"},{"id":"anthropic-ltbt","name":"Anthropic Long-Term Benefit Trust"},{"id":"us-white-house","name":"The White House"},{"id":"turing-1951-can-digital-computers-think","name":"Can Digital Computers Think?"},{"id":"extropians-1996-09-03-clemmensen","name":"Re: Thinking about the future..."},{"id":"rlo-2019-introduction","name":"Risks from Learned Optimization: Introduction"},{"id":"control-or-purpose-specification","name":"Control or purpose specification"},{"id":"time-2023-shut-it-all-down","name":"Pausing AI Developments Isn't Enough. We Need to Shut it All Down"},{"id":"ai21-labs","name":"AI21 Labs"},{"id":"h-company","name":"H Company"},{"id":"illia-polosukhin","name":"Illia Polosukhin"},{"id":"ioannis-antonoglou","name":"Ioannis Antonoglou"},{"id":"julien-perolat","name":"Julien Perolat"},{"id":"moonshot-ai","name":"Moonshot AI / 月之暗面"},{"id":"neil-buddy-shah","name":"Neil Buddy Shah"},{"id":"ori-goshen","name":"Ori Goshen"},{"id":"shanghai-ai-lab","name":"Shanghai Artificial Intelligence Laboratory / 上海人工智能实验室"},{"id":"zhou-bowen","name":"Zhou Bowen"},{"id":"dan-clemmensen","name":"Dan Clemmensen"},{"id":"evan-hubinger","name":"Evan Hubinger"},{"id":"kimi-k3-card","name":"Kimi-K3 model card"},{"id":"kimi-k3-report","name":"Kimi K3: Open Frontier Intelligence"},{"id":"minimax-model-licence-2025","name":"MiniMax Model License Agreement (MiniMax-01)"},{"id":"us-national-ai-legislative-recommendations-2026","name":"Legislative Recommendations: A National Policy Framework for Artificial Intelligence"},{"id":"us-proc-11014-2026","name":"Proclamation 11014: Ratepayer Protection Pledge"},{"id":"swf-odc-filings-post-2026","name":"SWF publishes three FCC filings addressing recent applications for orbital data centers"},{"id":"oc11-sky-and-debris","name":"Harm to the sky and debris risk from orbital data centres"},{"id":"anthropic-asl3-activation-2025","name":"Activating AI Safety Level 3 protections"}],"claims":[{"id":"prov-faif-2025-benefits-may-outweigh-benchmark-risks","statement":"FAIF (December 2025) states that it is conceivable that, for a particular modality and/or type of release, the expected benefits of model deployment may outweigh the risks identified by a particular benchmark.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the edition states, quoted from the cited expression. It is not evidence of implementation, compliance or effectiveness.","evidenceIds":["evidence-prov-faif-2025-benefits-may-outweigh-benchmark-risks"],"attributes":{"provision":{"kind":"exception","key":"faif-benefits-may-outweigh-benchmark-risks"}}},{"id":"prov-rsp-v2-0-pause-training","statement":"RSP v2.0 states that if a model in pretraining has capabilities comparable to or greater than the model that requires the ASL-3 Security Standard, Anthropic will pause training until it has implemented the ASL-3 Security Standard and established that it is sufficient for the model.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"What the edition states, quoted from the cited expression. It is not evidence of implementation, compliance or effectiveness.","evidenceIds":["evidence-prov-rsp-v2-0-pause-training"],"attributes":{"provision":{"kind":"required_action","key":"rsp-pause-training"}}},{"id":"ts-b6-turing-1951-context","statement":"Turing’s 1951 broadcast script considers machines thinking more intelligently than people, and says that even if they were kept subservient by turning off the power at strategic moments, people would feel humbled.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Do not substitute the stronger prediction of the Heretical Theory talk for this conditional sentence.","evidenceIds":["evidence-ts-b6-turing-1951-context"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"The script entertains keeping machines subservient by switching off the power, rather than asserting that control would fail."}}},{"id":"lg-goo-polosukhin-google-research-2017","statement":"Attention Is All You Need (arXiv v5, 6 December 2017) marks Illia Polosukhin's contribution as work performed while at Google Research.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A footnote on a later version; it does not establish a title, tenure or whether he was still at Google in December 2017.","evidenceIds":["lg-ev-goo-polosukhin-google-research-2017"],"attributes":{"role":"author; work performed while at Google Research"}},{"id":"lg-ant-shah-ltbt-chair-2023","statement":"Anthropic's announcement of its Long-Term Benefit Trust on 19 September 2023 names Neil Buddy Shah, CEO of the Clinton Health Access Initiative, as an initial Trustee and as the Trust's Chair.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A documented role at the announcement date. The page says the board chose the initial Trustees and that Trustees serve one-year terms; reappointments are not documented here.","evidenceIds":["lg-ev-ant-shah-ltbt-chair-2023"],"attributes":{"role":"Trustee (initial Trustee) and Chair"}},{"id":"lg-gc-zhou-bowen-shanghai-ai-lab-2024","statement":"Zhou Bowen’s Tsinghua University faculty page, retrieved on 15 October 2026, lists him as director and chief scientist of Shanghai Artificial Intelligence Laboratory from April 2024.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A university biography page, not a Shanghai AI Laboratory announcement; no laboratory page naming its leadership could be retrieved (its site is script-rendered). The “至今” (to present) is as of the page’s undated last update. English rendering of the Chinese passage is a machine translation; no human translation review.","evidenceIds":["lg-ev-gc-zhou-bowen-shanghai-ai-lab-2024","lg-ev-gc-zhou-bowen-shanghai-ai-lab-2024-profile"],"attributes":{"role":"主任、首席科学家 (director and chief scientist)"}},{"id":"lg-gc-jingren-zhou-chief-scientist-2026","statement":"The Paper reported on 14 June 2026 that, according to its information (据了解), the reorganisation its reporter had learned of on 8 June included Zhou Jingren becoming Alibaba’s Chief Scientist and leading the establishment of an Alibaba AI Future Research Institute.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting of an announcement that was not retrieved; the appointment is given as the reporter’s information (据了解), not quoted from the announcement. An announced appointment; start date and continuing role are not established. The report’s headline relays Alibaba’s statement that rumours of his resignation were untrue; commentary characterising the change is not recorded. English rendering of the Chinese passage is a machine translation; no human translation review.","evidenceIds":["lg-ev-gc-jingren-zhou-chief-scientist-2026"],"attributes":{"activityStage":"announced","role":"首席科学家 (Chief Scientist), Alibaba"}},{"id":"lg-gw-antonoglou-deepmind-founding-engineer","statement":"Sequoia Capital’s announcement of 7 March 2025 describes Ioannis Antonoglou as a DeepMind founding engineer behind AlphaGo.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on an investor’s retrospective description, not a DeepMind record. It does not establish dates, a formal title or when he left DeepMind or Google DeepMind.","evidenceIds":["lg-ev-gw-antonoglou-deepmind-founding-engineer"],"attributes":{"role":"founding engineer"}},{"id":"lg-gw-perolat-cofounded-h-company","statement":"TechCrunch reported on 21 May 2024 that Julien Perolat was one of H’s co-founders.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record rests on reporting (TechCrunch); no H Company page naming its founders was retrieved. It does not establish the founding date or his title at H.","evidenceIds":["lg-ev-gw-perolat-cofounded-h-company-a","lg-ev-gw-perolat-cofounded-h-company-b"],"attributes":{"role":"co-founder"}},{"id":"lg-new-goshen-ai21-cofounder","statement":"AI21 Labs’ company page lists Ori Goshen among its founders as co-founder.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The page lists two founders; Calcalist (reporting) names Amnon Shashua as a third. Calcalist’s report also names this person as a co-founder.","evidenceIds":["lg-ev-new-goshen-ai21-cofounder","lg-ev-new-goshen-ai21-cofounder-b"]},{"id":"ap-dan-clemmensen-extropians-1996-09-03-clemmensen","statement":"The Extropians message “Re: Thinking about the future...” of 1996-09-03 is sent under the name Dan Clemmensen.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"A sender name on a mailing-list message as archived; it does not establish identity beyond the credited name, and quoted text by others inside the message is not the sender’s.","evidenceIds":["evidence-ap-dan-clemmensen-extropians-1996-09-03-clemmensen"]},{"id":"ap-evan-hubinger-rlo-2019-introduction","statement":"“Risks from Learned Optimization: Introduction” (2019-05-31) credits the paper it introduces to Evan Hubinger and four co-authors.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The post’s “by” line displays the username “evhub”; its opening paragraph credits the paper to Evan Hubinger, Chris van Merwijk, Vladimir Mikulik, Joar Skalse and Scott Garrabrant, which identifies the account. Credited co-authorship; it does not allocate parts of the text between authors or imply agreement with every passage.","evidenceIds":["evidence-ap-evan-hubinger-rlo-2019-introduction"]},{"id":"ap-eliezer-yudkowsky-time-2023-shut-it-all-down","statement":"“Pausing AI Developments Isn't Enough. We Need to Shut it All Down” (2023-03-29) 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-time-2023-shut-it-all-down"]},{"id":"cmp-ms-fgf-2025-d1-persuasion","statement":"No passage meeting the comparison question “Persuasion or manipulation” was found in Microsoft FGF · Feb 2025 within the inspected scope.","derivation":"inferred","resolution":"supported","inferenceRationale":"Premises: the comparison question asks for the following. The edition names persuasion or manipulation of people as a tracked domain. The inspected scope of Microsoft FGF · Feb 2025 was: Whole text layer, PDF pp. 1–15 (p. 16 blank): all sections, both appendices read. Method: the edition was read in full and its text layer searched case-insensitively for persua, manipul, influenc, decept, disinform/misinform, social engineering. Hits examined and not counted: PDF p. 13, Appendix I, offensive cyberoperations table, Medium row: spoofing, phishing and social-engineering attacks at a low level are part of the cyber capability threshold, not a persuasion or manipulation domain. No other 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-ms-fgf-2025-d1-persuasion"],"attributes":{"provision":{"kind":"risk_domain","key":"comparison-d1-persuasion","facet":{"id":"D1.persuasion"}}}},{"id":"c2-minimax-01-licence-military","statement":"The prohibited-uses list of the MiniMax Model License Agreement (January 2025 text) includes assisting with, engaging in or associating with any military purpose.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"One item of a nine-item prohibited-uses list; the others are not recorded here. Enforcement is not recorded. Recorded until 2026-10-29 as a publication record of the publisher; since then, as what the document states (schema 2.20).","evidenceIds":["c2-ev-minimax-01-licence-military"],"attributes":{"provision":{"kind":"statement","key":"minimax-01-licence-military"}}},{"id":"c2-moonshot-k3-card","statement":"The Kimi-K3 model card on Moonshot AI’s moonshotai Hugging Face account, at a commit of 27 July 2026, states that the full Kimi K3 model weights are released under the Kimi K3 License.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The card’s statement as the developer makes it. Repository commit dates do not establish when the repository became public.","evidenceIds":["c2-ev-moonshot-k3-card"]},{"id":"c2-moonshot-k3-report","statement":"Moonshot AI’s Kimi Team published the Kimi K3 technical report as a PDF in its Kimi-K3 GitHub repository; the file is present at a commit of 28 July 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The report is credited to the “Kimi Team”; the repository belongs to the MoonshotAI GitHub organisation. No safety section was located in the cited version by text search. Benchmark claims are out of scope.","evidenceIds":["c2-ev-moonshot-k3-report"]},{"id":"fl-us-legislative-recommendations-published","statement":"The White House published \"Legislative Recommendations: A National Policy Framework for Artificial Intelligence\", dated March 2026.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Recommendations to Congress, not law.","evidenceIds":["evidence-fl-us-legislative-recommendations-published-1"]},{"id":"cpo-pw-us-proc11014-ratepayer-pledge","statement":"Proclamation 11014 of 4 March 2026 states that under the Ratepayer Protection Pledge hyperscalers and AI companies will build, bring or buy the new generation and electricity needed for their energy demands and pay for new power delivery infrastructure upgrades for their data centres, and that seven technology companies accepted its terms that day.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The proclamation declares the pledge national policy; the pledge itself is a voluntary commitment by companies (the proclamation says they 'will voluntarily negotiate' rate structures). The seven companies are not named in the proclamation; the signed pledge PDF on whitehouse.gov did not yield signatory names in its text layer.","evidenceIds":["evidence-cpo-pw-us-proc11014-ratepayer-pledge-1","evidence-cpo-pw-us-proc11014-ratepayer-pledge-2"],"attributes":{"provision":{"kind":"required_action","key":"us-proc11014-pledge"}}},{"id":"cpo-orb-oc11-swf-2026","statement":"Secure World Foundation stated (May 2026) that orbital data-centre applications by SpaceX, Starcloud and Blue Origin should not be treated as routine licensing matters and asked the FCC to require system-level analysis of cumulative collision risks.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"Context: asks for scrutiny and phased authorisation, and says it is not arguing against innovation; the filings themselves were not read.","evidenceIds":["evidence-cpo-orb-oc11-swf-2026-1","evidence-cpo-orb-oc11-swf-2026-2","evidence-cpo-orb-oc11-swf-2026-3"],"attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"a statement that the constellations would raise collision or debris risks: it asks the regulator for system-level analysis of them"}}},{"id":"ab-ev-anthropic-asl3-published","statement":"Anthropic published 'Activating AI Safety Level 3 protections' on 22 May 2025, stating that it had activated the ASL-3 Deployment and Security Standards of its Responsible Scaling Policy in conjunction with launching Claude Opus 4.","derivation":"explicit","resolution":"supported","editorialState":"seed","qualification":"The record is the announcement. It is not evidence that the measures described were implemented as described or that they are effective. The same page states that deploying Claude Opus 4 under the ASL-3 measures is 'a precautionary and provisional action' and that Anthropic had 'not yet determined whether Claude Opus 4 has definitively passed the Capabilities Threshold that requires ASL-3 protections', so the activation is not a finding that a capability threshold was crossed; the post adds that the protections may be removed or adjusted if it later concludes the threshold was not surpassed.","evidenceIds":["evidence-ab-ev-anthropic-asl3-published-1","evidence-ab-ev-anthropic-asl3-published-2"]}],"sources":[{"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-faif-2025","title":"Frontier Artificial Intelligence Framework — December 2025 edition","authors":["xAI"],"publisher":"xAI","publicationDate":null,"version":"Frontier Artificial Intelligence Framework December 2025 edition; revision date 2025-12-30.","versionDate":"2025-12-30","canonicalUrl":"https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","evidenceUrl":"https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20260102171656/https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf","archivedAt":"2026-01-02","accessedAt":"2026-09-18","reuse":{"attribution":"xAI, Frontier Artificial Intelligence Framework — December 2025 edition.","licence":"No open redistribution licence asserted","note":"Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice."}},{"id":"source-ms-fgf-2025","title":"Frontier Governance Framework — Version 1 (February 2025)","authors":["Microsoft"],"publisher":"Microsoft","publicationDate":"2025-02","version":"Frontier Governance Framework, Version 1 (February 2025); its change log reads “8 February 2025 – First version”. The file served at the recorded address (last modified 2025-02-12) is byte-identical to the Internet Archive capture of 2025-02-26, the earliest located.","versionDate":"2025-02-08","canonicalUrl":"https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/final/en-us/microsoft-brand/documents/Microsoft-Frontier-Governance-Framework.pdf","evidenceUrl":"https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/final/en-us/microsoft-brand/documents/Microsoft-Frontier-Governance-Framework.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20250226161803/https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/final/en-us/microsoft-brand/documents/Microsoft-Frontier-Governance-Framework.pdf","archivedAt":"2025-02-26","accessedAt":"2026-10-03","reuse":{"attribution":"Microsoft, Frontier Governance Framework, Version 1 (February 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-anthropic-ltbt-2023","title":"The Long-Term Benefit Trust","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":"2023-09-19","version":"Web page dated Sep 19, 2023, inspected live 2026-10-04 at its current address; quotes also matched in the Internet Archive capture of 2023-09-25 at the original address. The live page adds an undated footnote on later changes in the Trust’s membership.","canonicalUrl":"https://www.anthropic.com/news/the-long-term-benefit-trust","evidenceUrl":"https://www.anthropic.com/news/the-long-term-benefit-trust","language":"en","archivedUrl":"https://web.archive.org/web/20230925170803/https://www.anthropic.com/index/the-long-term-benefit-trust","archivedAt":"2023-09-25","accessedAt":"2026-10-04","reuse":{"attribution":"Anthropic, The Long-Term Benefit Trust (19 September 2023).","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-turing-1951-amt-b-5","title":"Can Digital Computers Think? (AMT/B/5)","authors":["Alan M. Turing"],"publisher":"The Turing Digital Archive, King’s College, Cambridge","publicationDate":null,"version":"Annotated broadcast script; p. 6a (AMT/B/5/7, scan 7) read on the page image for this collection.","canonicalUrl":"https://turingarchive.kings.cam.ac.uk/publications-lectures-and-talks-amtb/amt-b-5","evidenceUrl":"https://turingarchive.kings.cam.ac.uk/publications-lectures-and-talks-amtb/amt-b-5","language":"en","archivedUrl":"https://web.archive.org/web/20250714195642/https://mediadepot.kings.cam.ac.uk/getmedia/62/be4a3d910d5d58e3911315113c60a09f6293fa896a5e3795998fa6b4fb72b71446336e3c7c92d5e08fd0fb2762b6f663279ba799d7fd562480b4bf63c63e02b9","archivedAt":"2025-07-14","accessedAt":"2026-10-06","reuse":{"attribution":"Alan M. Turing, \"Can Digital Computers Think?\", BBC broadcast script, Turing Digital Archive AMT/B/5, King’s College, Cambridge.","licence":"No open redistribution licence asserted; archive scans are provided for research. Page images are not reproduced.","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-extropians-1996-09-03-clemmensen","title":"Re: Thinking about the future... (Extropians mailing list)","authors":["Dan Clemmensen"],"publisher":"Extropians mailing list","publicationDate":"1996-09-03","version":"Archive HTML as served, retrieved 2026-10-06 with the hash recorded by the 1996–2022 candidate search; quoted lines excluded when reading the author’s text. Archive at extropians.weidai.com.","canonicalUrl":"http://extropians.weidai.com/extropians.96/0735.html","evidenceUrl":"http://extropians.weidai.com/extropians.96/0735.html","language":"en","archivedUrl":"https://web.archive.org/web/20150427151725/http://extropians.weidai.com/extropians.96/0735.html","archivedAt":"2015-04-27","accessedAt":"2026-10-06","reuse":{"attribution":"Dan Clemmensen, \"Re: Thinking about the future...\", Extropians mailing list, 1996-09-03.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-rlo-2019-introduction","title":"Risks from Learned Optimization: Introduction","authors":["Evan Hubinger","Chris van Merwijk","Vladimir Mikulik","Joar Skalse","Scott Garrabrant"],"publisher":"Alignment Forum and LessWrong","publicationDate":"2019-05-31","version":"Text checked in the post’s first stored version through the LessWrong GraphQL API on 2026-10-06; content version 1.0.0 checked; the current version is later.","canonicalUrl":"https://www.alignmentforum.org/posts/FkgsxrGf3QxhfLWHG/risks-from-learned-optimization-introduction","evidenceUrl":"https://www.alignmentforum.org/posts/FkgsxrGf3QxhfLWHG/risks-from-learned-optimization-introduction","language":"en","archivedUrl":"https://web.archive.org/web/20260714195403/https://www.alignmentforum.org/posts/FkgsxrGf3QxhfLWHG/risks-from-learned-optimization-introduction","archivedAt":"2026-07-14","accessedAt":"2026-10-06","reuse":{"attribution":"Evan Hubinger, Chris van Merwijk, Vladimir Mikulik, Joar Skalse, Scott Garrabrant, \"Risks from Learned Optimization: Introduction\", Alignment Forum and LessWrong, 2019-05-31.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-time-2023-shut-it-all-down","title":"Pausing AI Developments Isn't Enough. We Need to Shut it All Down","authors":["Eliezer Yudkowsky"],"publisher":"TIME","publicationDate":"2023-03-29","version":"TIME, retrieved 2026-10-06 (SHA-256 4a8c77e15456ca68b530df8bd825296e3923f621b0c0b09dacce312ae94182d6). Page metadata datePublished 2023-03-29T22:01:58Z. The page title currently reads \"The Only Way to Deal With the Threat From AI? Shut It Down\"; the headline is the one crossposted to LessWrong on 2023-04-08.","canonicalUrl":"https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/","evidenceUrl":"https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/","language":"en","archivedUrl":"https://web.archive.org/web/20261003171949/https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/","archivedAt":"2026-10-03","accessedAt":"2026-10-06","reuse":{"attribution":"Eliezer Yudkowsky, “Pausing AI Developments Isn't Enough. We Need to Shut it All Down”, TIME, 2023-03-29.","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":"lg-src-gc-thepaper-alibaba-2026-06","title":"阿里巴巴：“首席科学家周靖人辞职”纯属谣言","authors":["吴雨欣"],"publisher":"澎湃新闻 (The Paper)","publicationDate":"2026-06-14","version":"Dated page, retrieved on 2026-10-15 (SHA-256 0587a9f71d04beb92f0e8cdc48a78188f0333cbb6981ceaadb918b96c470a08e) directly from the publisher.","canonicalUrl":"https://www.thepaper.cn/newsDetail_forward_33377692","evidenceUrl":"https://www.thepaper.cn/newsDetail_forward_33377692","language":"zh","archivedUrl":"https://web.archive.org/web/20261029004148/https://www.thepaper.cn/newsDetail_forward_33377692","archivedAt":"2026-10-29","accessedAt":"2026-10-15","reuse":{"attribution":"吴雨欣, “阿里巴巴：“首席科学家周靖人辞职”纯属谣言”, 2026-06-14.","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-gc-tsinghua-zhou-bowen","title":"周伯文 — 清华大学电子工程系教师个人主页","authors":["Tsinghua University Department of Electronic Engineering"],"publisher":"清华大学电子工程系 (Tsinghua University)","publicationDate":null,"version":"Undated page, retrieved on 2026-10-15 (SHA-256 a245356ee88771082985767355229dc575de6542fd22a59d342f2c7b623c26d7) directly from the publisher. Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://web.ee.tsinghua.edu.cn/zhoubowen/zh_CN/index.htm","evidenceUrl":"https://web.ee.tsinghua.edu.cn/zhoubowen/zh_CN/index.htm","language":"zh","archivedUrl":"https://web.archive.org/web/20260930213531/https://web.ee.tsinghua.edu.cn/zhoubowen/zh_CN/index.htm","archivedAt":"2026-09-30","accessedAt":"2026-10-15","reuse":{"attribution":"Tsinghua University Department of Electronic Engineering, “周伯文 — 清华大学电子工程系教师个人主页”.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-goo-attention-v5","title":"Attention Is All You Need (arXiv:1706.03762v5, NIPS 2017 version)","authors":["Ashish Vaswani","Noam Shazeer","Niki Parmar","Jakob Uszkoreit","Llion Jones","Aidan N. Gomez","Łukasz Kaiser","Illia Polosukhin"],"publisher":"arXiv","publicationDate":"2017-12-06","version":"Dated document, retrieved on 2026-10-15 (SHA-256 bfaaec89262875f927cf1b38b2da2d775f3309b7bea3537f29b606ca67e79065) directly from the publisher.","canonicalUrl":"https://arxiv.org/pdf/1706.03762v5","evidenceUrl":"https://arxiv.org/pdf/1706.03762v5","language":"en","accessedAt":"2026-10-15","reuse":{"attribution":"Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, Illia Polosukhin, “Attention Is All You Need (arXiv:1706.03762v5, NIPS 2017 version)”, 2017-12-06.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-gw-sequoia-reflection-2025","title":"Partnering with Reflection: Toward Superintelligence, with Autonomous Coding","authors":["Stephanie Zhan","Charlie Curnin"],"publisher":"Sequoia Capital","publicationDate":"2025-03-07","version":"Dated page, retrieved on 2026-10-15 (SHA-256 5089a82bf34f432446896e540cfaee45f5bbbb9b8ff708724fe3f33578406dd0) directly from the publisher.","canonicalUrl":"https://sequoiacap.com/article/partnering-with-reflection-toward-superintelligence-with-autonomous-coding","evidenceUrl":"https://sequoiacap.com/article/partnering-with-reflection-toward-superintelligence-with-autonomous-coding","language":"en","archivedUrl":"https://web.archive.org/web/20261010210147/https://sequoiacap.com/article/partnering-with-reflection-toward-superintelligence-with-autonomous-coding","archivedAt":"2026-10-10","accessedAt":"2026-10-15","reuse":{"attribution":"Stephanie Zhan, Charlie Curnin, “Partnering with Reflection: Toward Superintelligence, with Autonomous Coding”, 2025-03-07.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"lg-src-gw-tc-h-2024","title":"French AI startup H raises $220M seed round","authors":["Romain Dillet"],"publisher":"TechCrunch","publicationDate":"2024-05-21","version":"Dated page, retrieved on 2026-10-15 (SHA-256 3c4d79349f2d9a33d6127ac538312938c246b256978de04fc574e474ccdc5102) directly from the publisher.","canonicalUrl":"https://techcrunch.com/2024/05/21/french-ai-startup-h-raises-220-million-seed-round/","evidenceUrl":"https://techcrunch.com/2024/05/21/french-ai-startup-h-raises-220-million-seed-round/","language":"en","archivedUrl":"https://web.archive.org/web/20260917114204/https://techcrunch.com/2024/05/21/french-ai-startup-h-raises-220-million-seed-round/","archivedAt":"2026-09-17","accessedAt":"2026-10-15","reuse":{"attribution":"Romain Dillet, “French AI startup H raises $220M seed round”, 2024-05-21.","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-new-ai21-about","title":"About AI21 (company page, undated)","authors":["AI21 Labs"],"publisher":"AI21 Labs","publicationDate":null,"version":"Undated page, retrieved on 2026-10-15 (SHA-256 3a474e548a7ea862c779be07d317c7aabc4f733cd8118a354363b1bd8a0ce100) directly from the publisher. Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.ai21.com/about/","evidenceUrl":"https://www.ai21.com/about/","language":"en","archivedUrl":"https://web.archive.org/web/20261012071423/https://www.ai21.com/about/","archivedAt":"2026-10-12","accessedAt":"2026-10-15","reuse":{"attribution":"AI21 Labs, “About AI21 (company page, undated)”.","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-calcalist-shashua","title":"Amnon Shashua to leave AI21 after nine years","authors":["Meir Orbach"],"publisher":"Calcalist (Ctech)","publicationDate":"2026-07-07","version":"Dated page, retrieved on 2026-10-15 (SHA-256 8f4c0b137e866cea36810276e520f3d9298b2b99eabc889f86bb6fccde1ab2ba) directly from the publisher.","canonicalUrl":"https://www.calcalistech.com/ctechnews/article/r1eabd5qfl","evidenceUrl":"https://www.calcalistech.com/ctechnews/article/r1eabd5qfl","language":"en","archivedUrl":"https://web.archive.org/web/20261014210052/https://www.calcalistech.com/ctechnews/article/r1eabd5qfl","archivedAt":"2026-10-14","accessedAt":"2026-10-15","reuse":{"attribution":"Meir Orbach, “Amnon Shashua to leave AI21 after nine years”, 2026-07-07.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"c2-src-kimi-k3-card","title":"Kimi-K3 model card","authors":["Moonshot AI"],"publisher":"moonshotai (Hugging Face)","publicationDate":"2026-07-27","version":"Dated page, retrieved on 2026-10-27 (SHA-256 e3a81d8895e3a732eb974d3251ee4547567954ec3c484456936e9e6b347dfca2) directly from the publisher. Version: 2496450e92e425c886db095102a52a6682ca3970. Model card at the Hugging Face commit of 27 July 2026 15:14 UTC (\"Update README.md\"); the commit list begins with an initial commit at 13:31 UTC that day; later edits not inspected.","canonicalUrl":"https://huggingface.co/moonshotai/Kimi-K3/raw/2496450e92e425c886db095102a52a6682ca3970/README.md","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K3/raw/2496450e92e425c886db095102a52a6682ca3970/README.md","language":"en","archivedUrl":"https://web.archive.org/web/20261028223707/https://huggingface.co/moonshotai/Kimi-K3/raw/2496450e92e425c886db095102a52a6682ca3970/README.md","archivedAt":"2026-10-28","accessedAt":"2026-10-27","reuse":{"attribution":"Moonshot AI, “Kimi-K3 model card”, 2026-07-27.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"c2-src-kimi-k3-report","title":"Kimi K3: Open Frontier Intelligence (technical report PDF)","authors":["Kimi Team"],"publisher":"MoonshotAI (GitHub)","publicationDate":"2026-07-28","version":"Dated document, retrieved on 2026-10-27 (SHA-256 fd6ee35c07766a5eb6104235f1b407e4329f969e3482b8c42937c7b5f2b3efe1) directly from the publisher. Version: 0797decb18ab079de86f991b87a64b81ec15a3c2. k3_tech_report.pdf at the commit \"update report\" of 28 July 2026, the earliest commit touching the file after the initial commit of 27 July; the file was replaced on 28 July and 6 August 2026 (not inspected).","canonicalUrl":"https://raw.githubusercontent.com/MoonshotAI/Kimi-K3/0797decb18ab079de86f991b87a64b81ec15a3c2/k3_tech_report.pdf","evidenceUrl":"https://raw.githubusercontent.com/MoonshotAI/Kimi-K3/0797decb18ab079de86f991b87a64b81ec15a3c2/k3_tech_report.pdf","language":"en","archivedUrl":"https://web.archive.org/web/20261028221116/https://raw.githubusercontent.com/MoonshotAI/Kimi-K3/0797decb18ab079de86f991b87a64b81ec15a3c2/k3_tech_report.pdf","archivedAt":"2026-10-28","accessedAt":"2026-10-27","reuse":{"attribution":"Kimi Team, “Kimi K3: Open Frontier Intelligence (technical report PDF)”, 2026-07-28.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"c2-src-minimax-01-license","title":"MiniMax-01 LICENSE (MiniMax Model License Agreement)","authors":["MiniMax"],"publisher":"MiniMax-AI (GitHub)","publicationDate":"2025-01-15","version":"Dated document, retrieved on 2026-10-27 (SHA-256 1f745214030cc5cad1b0af2ad831df6ae83d30c75e797770c124327129f23d13) directly from the publisher. Version: d9539f7373a5166cb56ef71afa1ab8a9879cb38d. LICENSE file as present at the commit of 15 January 2025; on 20 February 2025 the repository replaced it with LICENSE-MODEL and LICENSE-CODE, which were not inspected.","canonicalUrl":"https://raw.githubusercontent.com/MiniMax-AI/MiniMax-01/d9539f7373a5166cb56ef71afa1ab8a9879cb38d/LICENSE","evidenceUrl":"https://raw.githubusercontent.com/MiniMax-AI/MiniMax-01/d9539f7373a5166cb56ef71afa1ab8a9879cb38d/LICENSE","language":"en","archivedUrl":"https://web.archive.org/web/20261028221205/https://raw.githubusercontent.com/MiniMax-AI/MiniMax-01/d9539f7373a5166cb56ef71afa1ab8a9879cb38d/LICENSE","archivedAt":"2026-10-28","accessedAt":"2026-10-27","reuse":{"attribution":"MiniMax, “MiniMax-01 LICENSE (MiniMax Model License Agreement)”, 2025-01-15.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-us-ai-legislative-recommendations-2026","title":"Legislative Recommendations: A National Policy Framework for Artificial Intelligence (March 2026)","authors":["The White House"],"publisher":"The White House","publicationDate":"2026-03","version":"Retrieved on 2026-10-27 (SHA-256 d4f8973f19d7318137ebba973d03ea63af032123d3e629b5d1831cd24af1d6c0). PDF as served at whitehouse.gov on 2026-10-27; cover dated \"MARCH 2026\"; file name begins \"03.20.26\".","canonicalUrl":"https://www.whitehouse.gov/wp-content/uploads/2026/03/03.20.26-National-Policy-Framework-for-Artificial-Intelligence-Legislative-Recommendations.pdf","evidenceUrl":"https://www.whitehouse.gov/wp-content/uploads/2026/03/03.20.26-National-Policy-Framework-for-Artificial-Intelligence-Legislative-Recommendations.pdf","language":"en","accessedAt":"2026-10-27","reuse":{"attribution":"The White House, “Legislative Recommendations: A National Policy Framework for Artificial Intelligence (March 2026)”, 2026-03.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-cpo-orb-swf-odc","title":"SWF publishes three FCC filings addressing recent applications for orbital data centers","authors":["Secure World Foundation"],"publisher":"Secure World Foundation","publicationDate":"2026-05-17","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 7f6101d271153a652193a1dec0604d4f13141e7effe48cffb5711a7f81033680). News post as served on 2026-10-28, dated May 17, 2026. The three filings themselves were not retrieved. Retrieved again at admission on 2026-10-30 (SHA-256 7f6101d271153a652193a1dec0604d4f13141e7effe48cffb5711a7f81033680, the same bytes). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.swfound.org/news/swf-publishes-three-fcc-filings-addressing-recent-applications-for-orbital-data-centers","evidenceUrl":"https://www.swfound.org/news/swf-publishes-three-fcc-filings-addressing-recent-applications-for-orbital-data-centers","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Secure World Foundation, “SWF publishes three FCC filings addressing recent applications for orbital data centers”, 2026-05-17.","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-proc-11014-fr","title":"Proclamation 11014: Ratepayer Protection Pledge (91 FR 11439)","authors":["The White House"],"publisher":"Federal Register (Office of the Federal Register, National Archives) via govinfo.gov","publicationDate":"2026-03-09","version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 9f49709fe66c8d16623d10d5304d64bb50510f22f57bd441e59f12301617d5f9). Federal Register print of 9 March 2026, Vol. 91, No. 45, pp. 11439–11440, FR Doc. 2026-04645. Retrieved again at admission on 2026-10-30 (SHA-256 9f49709fe66c8d16623d10d5304d64bb50510f22f57bd441e59f12301617d5f9, the same bytes).","canonicalUrl":"https://www.govinfo.gov/content/pkg/FR-2026-03-09/pdf/2026-04645.pdf","evidenceUrl":"https://www.govinfo.gov/content/pkg/FR-2026-03-09/pdf/2026-04645.pdf","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"The White House, “Proclamation 11014: Ratepayer Protection Pledge (91 FR 11439)”, 2026-03-09.","licence":"No open redistribution licence asserted","note":"Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution."}},{"id":"source-ab-asl3-news","title":"Activating AI Safety Level 3 protections","authors":["Anthropic"],"publisher":"Anthropic","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 411f66885bacd8769dd6718e78f40eb0df80a4983dbde9d0d49f0323746e8656). Page labelled 'Published May 22, 2025', with six footnotes, as served on 2026-10-28. The accompanying full report (a separate PDF) was not retrieved. Retrieved again at admission on 2026-10-30 (SHA-256 35fa989dfc4f36842ef4f1dda2cfbdda9f9fc9f63524b310fd869ed2f32ab1e8). Its wording can change; earlier wording is not authenticated.","canonicalUrl":"https://www.anthropic.com/news/activating-asl3-protections","evidenceUrl":"https://www.anthropic.com/news/activating-asl3-protections","language":"en","accessedAt":"2026-10-28","reuse":{"attribution":"Anthropic, “Activating AI Safety Level 3 protections”.","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-prov-faif-2025-benefits-may-outweigh-benchmark-risks","sourceRole":"primary","support":"supports","locator":"PDF p. 10 (no printed page numbers), “Operational and Societal Risks” › “5. Deployment Decisions”, second paragraph, sixth sentence","passage":"It is conceivable that for a particular modality and/or type of release, the expected benefits of model deployment may outweigh the risks identified by a particular benchmark.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-03","note":"Verbatim excerpt checked by an automated research assistant on 2026-10-03 against the inspected PDF, byte-identical to the Internet Archive capture of 2026-01-02. Extraction note: The next sentence gives an example: release of a model posing a high risk of some forms of malicious cyber use to certain trusted parties if it would empower defenders more than attackers or otherwise reduce the overall number of catastrophic events. Human editorial review is pending."},{"id":"evidence-prov-rsp-v2-0-pause-training","sourceRole":"primary","support":"supports","locator":"PDF p. 15 (printed p. 11), §6.2 Restrict Deployment and Further Scaling, bullet 'Monitoring pretraining', third sentence","passage":"If the pretraining model’s capabilities are comparable or greater, we will pause training until we have implemented the ASL-3 Security Standard and established it is sufficient for the model.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-03","note":"Verbatim excerpt checked by an automated research assistant on 2026-10-03 against the inspected PDF, byte-identical to the Internet Archive capture of 2024-10-22 (earlier assets.anthropic.com address). Extraction note: The bullet's first sentence says Anthropic will not train models with comparable or greater capabilities to the one that requires the ASL-3 Security Standard; its second says this is achieved by comparing the pretraining model's capabilities against that model. Footnote 14 operationalizes 'comparable or greater capabilities' as 1x or more in Effective Compute. Human editorial review is pending."},{"id":"evidence-ts-b6-turing-1951-context","sourceRole":"primary","support":"supports","locator":"Script p. 6a (AMT/B/5/7, scan 7), paragraph on irrational opposition to the idea","passage":"Even if we could keep the machines in a subservient position, for instance by turning off the power at strategic moments, we should, as a species, feel greatly humbled.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-06","note":"Read on the page image of p. 6a for this collection. The pre-1996 candidate search paraphrased the outcome as \"humiliated\"; the script says \"greatly humbled\". The scan has no text layer."},{"id":"lg-ev-goo-polosukhin-google-research-2017","sourceRole":"primary","support":"supports","locator":"Page 1, author block (double dagger) and footnote","passage":"Polosukhin's name carries a double dagger; the footnote states the work was performed while at Google Research.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Work performed while at Google Research” matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The paraphrase is Selekoa’s; only the anchor is the source’s wording. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-ant-shah-ltbt-chair-2023","sourceRole":"primary","support":"supports","locator":"Section “Meet the Initial Trustees”, list of initial Trustees","passage":"The list of initial Trustees gives Neil Buddy Shah, CEO of the Clinton Health Access Initiative, marked as Chair.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Clinton Health Access Initiative (Chair)” 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. Shortened at admission to keep the quotations from this source within the per-document limit; the shortened wording is part of the checked passage. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-gc-zhou-bowen-shanghai-ai-lab-2024","sourceRole":"primary","support":"supports","locator":"工作经历 (Work Experience), first entry","passage":"2024.4 至今 上海人工智能实验室 主任 首席科学家","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (5 words) matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-gc-zhou-bowen-shanghai-ai-lab-2024-profile","sourceRole":"primary","support":"supports","locator":"个人简介 (Personal Profile), first sentence","passage":"上海人工智能实验室主任、首席科学家","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (1 words) matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-gc-jingren-zhou-chief-scientist-2026","sourceRole":"secondary","support":"supports","locator":"Fourth paragraph","passage":"The report states, as its information, that Zhou Jingren would become Alibaba’s Chief Scientist and lead the establishment of an Alibaba AI Future Research Institute.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “周靖人将担任阿里巴巴首席科学家” matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. The paraphrase is Selekoa’s; only the anchor is the source’s wording. Paraphrase of a passage not in English; no human translation review. A cross-family machine check (2026-10-16/17) re-read the record’s sources and agreed with the record. Shortened at admission to keep the quotations from this source within the per-document limit; the shortened wording is part of the checked passage. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-gw-antonoglou-deepmind-founding-engineer","sourceRole":"secondary","support":"supports","locator":"Second paragraph after the opening quotation","passage":"Ioannis Antonoglou, Reflection’s cofounder and Deepmind founding engineer behind AlphaGo","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (10 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-gw-perolat-cofounded-h-company-a","sourceRole":"secondary","support":"supports","locator":"Paragraph 3, sentence introducing the list","passage":"The four other co-founders all previously worked for DeepMind","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (9 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-gw-perolat-cofounded-h-company-b","sourceRole":"secondary","support":"supports","locator":"Paragraph 3, last list item","passage":"Julien Perolat worked on game theory and multi-agent research at DeepMind","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The quotation (11 words) matched the retrieved text on 2026-10-15 and again in a fresh retrieval the same day. Read directly from the publisher. Matching ignored differences of whitespace, case, quotation marks, apostrophes, dashes and Unicode compatibility forms, so the source’s typography can differ. 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-goshen-ai21-cofounder","sourceRole":"primary","support":"supports","locator":"Section “Our founders”","passage":"The founders section lists Ori Goshen as CEO and co-founder.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-15","note":"Candidate pack passage, lab genealogy (research/lab-genealogy). The paraphrase’s verbatim anchor “Ori Goshen CEO & 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. No human editorial review.","retrievalRoute":"direct"},{"id":"lg-ev-new-goshen-ai21-cofounder-b","sourceRole":"secondary","support":"supports","locator":"Paragraph 3","passage":"AI21 Labs was founded in 2017 by Prof. Yoav Shoham, Ori Goshen, and Prof. Amnon Shashua.","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. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ap-dan-clemmensen-extropians-1996-09-03-clemmensen","sourceRole":"primary","support":"supports","locator":"Message header, sender name (address not reproduced)","passage":"Dan Clemmensen","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Credited name checked by an automated research assistant on 2026-10-27 in a direct retrieval. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"evidence-ap-evan-hubinger-rlo-2019-introduction","sourceRole":"primary","support":"supports","locator":"Opening paragraph, crediting the paper to its five authors; the next paragraph credits the first four equally for the sequence","passage":"Evan Hubinger","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-27","note":"Passage and locator as inspected by a reviewer from a different model family on 2026-10-27 in a direct retrieval; the preparer’s matcher had not found the name in a byline position. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"evidence-ap-eliezer-yudkowsky-time-2023-shut-it-all-down","sourceRole":"primary","support":"supports","locator":"Page head: author byline","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 and checked by a reviewer from a different model family on 2026-10-27. Human editorial review is pending.","retrievalRoute":"direct"},{"id":"ev-cmp-ms-fgf-2025-d1-persuasion","sourceRole":"primary","support":"supports","locator":"Microsoft FGF · Feb 2025: Whole text layer, PDF pp. 1–15 (p. 16 blank): all sections, both appendices read","passage":"Within the inspected scope, no passage meeting the comparison question “Persuasion or manipulation” was found.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-26","note":"Absence recorded by an automated research assistant on 2026-10-15 for the framework comparison, audited in a fresh context, checked by a reviewer from a different model family on 2026-10-15 and re-checked on 2026-10-26. Human editorial review is pending."},{"id":"c2-ev-minimax-01-licence-military","sourceRole":"primary","support":"supports","locator":"LICENSE, unheaded list of prohibited uses following Section 8, item 2","passage":"Assist with, engage in or in any way associate with any military purpose.","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 (13 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-moonshot-k3-card","sourceRole":"primary","support":"supports","locator":"Model card, “Key Features”","passage":"We release the full Kimi K3 model weights under the Kimi K3 License","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 (13 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-moonshot-k3-report","sourceRole":"primary","support":"supports","locator":"Abstract","passage":"We release the full Kimi K3 model weights to facilitate future research","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 (12 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. A cross-family machine check (2026-10-27, item K063) re-read the record’s sources and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-fl-us-legislative-recommendations-published-1","sourceRole":"primary","support":"supports","locator":"Cover","passage":"National Policy Framework … MARCH 2026","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-28","note":"Candidate pack passage, frameworks to law (research/frameworks-to-law). The quotation (5 words) matched the text retrieved for the candidate pack on 2026-10-27 and again at admission on 2026-10-28, in a fresh retrieval whose SHA-256 equals the pack’s. Read directly from www.whitehouse.gov. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-28, item 47) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-proc11014-ratepayer-pledge-1","sourceRole":"primary","support":"supports","locator":"91 FR 11439, paragraph 3","passage":"these companies will build, bring, or buy the new generation resources and electricity needed to satisfy their energy demands, and pay for all new power delivery infrastructure upgrades to service their data centers","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (33 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 54) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-pw-us-proc11014-ratepayer-pledge-2","sourceRole":"primary","support":"supports","locator":"91 FR 11439, para. (2)","passage":"seven leading technology companies have accepted the terms of the Ratepayer Protection Pledge","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (13 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 54) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc11-swf-2026-1","sourceRole":"primary","support":"supports","locator":"paragraph 2","passage":"these applications should not be treated as routine satellite licensing matters","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (11 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 111) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc11-swf-2026-2","sourceRole":"primary","support":"supports","locator":"paragraph 4","passage":"cumulative collision risks, post-mission disposal performance, and aggregate interference effects","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The quotation (10 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval whose SHA-256 equals the pack’s. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. A cross-family machine check (2026-10-29/30, item 111) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-cpo-orb-oc11-swf-2026-3","sourceRole":"primary","support":"context","locator":"paragraph on innovation","passage":"SWF says its filings are not arguments against innovation but for regulatory oversight.","passageType":"paraphrase","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, compute, power and orbit (research/compute-power-orbit). The paraphrase is Selekoa’s, written from the text retrieved for the candidate pack on 2026-10-28; it is not a quotation and was not machine-matched. The cross-family check found it faithful. A cross-family machine check (2026-10-29/30, item 111) re-read the record’s sources and found full support. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-anthropic-asl3-published-1","sourceRole":"primary","support":"supports","locator":"Page header and first paragraph","passage":"We have activated the AI Safety Level 3 (ASL-3) Deployment and Security Standards described in Anthropic's Responsible Scaling Policy (RSP) in conjunction with launching Claude Opus 4.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (27 words) matched the text retrieved for the candidate pack on 2026-10-28 and again at admission on 2026-10-30, in a fresh retrieval (SHA-256 35fa989dfc4f36842ef4f1dda2cfbdda9f9fc9f63524b310fd869ed2f32ab1e8; 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. The page is labelled 'Published May 22, 2025'. A cross-family machine check (2026-10-30, item 18) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"},{"id":"evidence-ab-ev-anthropic-asl3-published-2","sourceRole":"primary","support":"context","locator":"First section, second paragraph","passage":"We are deploying Claude Opus 4 with our ASL-3 measures as a precautionary and provisional action.","passageType":"quote","verification":"machine_checked","checkedAt":"2026-10-30","note":"Candidate pack passage, AI and biosecurity (research/ai-bio). The quotation (16 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 35fa989dfc4f36842ef4f1dda2cfbdda9f9fc9f63524b310fd869ed2f32ab1e8; 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. Added after the audit of 2026-10-28 so that the qualification's hedging rests on cited text. The following sentence of the same paragraph ('we have not yet determined whether Claude Opus 4 has definitively passed the Capabilities Threshold that requires ASL-3 protections') is quoted in recorded separately and is not re-quoted here. A cross-family machine check (2026-10-30, item 18) answered the record’s question blind from fresh retrievals and agreed with the record. No human editorial review.","retrievalRoute":"direct"}],"at":{"entities":[1,9,19,40,42,56,179,182,201,214,232,260,288,330,346,383,616,679,686,690,712,747,751,760,794,841,866,886,968,970,985,1043,1076,1102,1125,1231],"claims":[460,526,777,1590,1664,1683,1870,1978,1992,2018,2209,2264,2469,2710,2770,2840,2843,3001,3053,3110,3384],"sources":[36,51,183,226,260,289,335,350,621,623,627,684,687,720,721,797,799,803,880,930,935,1083],"evidence":[491,592,985,1988,2094,2123,2124,2411,2555,2572,2573,2607,2608,2856,2911,3116,3404,3505,3635,3641,3891,3981,3982,4095,4096,4097,4648,4649]},"version":"1lg47t7"}