Public recordOpen fileA583F601

Public record case file · Latest recorded event Oct 2, 2026

Federal AI Safety Governance

Accountability file

A public-record case file for U.S. federal AI-safety governance and oversight: executive orders on AI, the NIST AI Risk Management Framework, the standards body launched as the U.S. AI Safety Institute and reconstituted as the Center for AI Standards and Innovation, FTC enforcement on AI claims, GAO oversight of federal AI use, and congressional AI hearings and legislation. It keeps executive-order status, standards guidance, enforcement actions, oversight reports, and legislative records separate until a public record supports the exact claim, and it does not rate AI model quality, capability, or safety.

Original editorial artwork generated for Hugin.

Latest recorded event: Oct 2, 2026
Source Anchors: 30 verified
Timeline: 24 events
Roles: 13 public-record
Fingerprint: A583 F601 EBF3

Submit source Data exports →

Saved reading0In this browser

Keep roles, events, and source records here to pick up later. Your list stays in this browser and can be lost if you clear site data. Copy your saved links to keep a backup in your notes.

Your reading list is empty. Use a save button on an event, source, or role dossier to add it here.

Start reading

24 events · 13 roles · 30 sources

Recent developments in the record

The latest dated events in this file. Open an event for its cited sources and evidence limits.

Oct 2, 2026
On the day EO 14434 prints, GSA's America.gov already writes 'SI', including in its link text for four outside articles whose web addresses read 'ai' →records release review · 4 source records
Read summary and limits
Read October 2, the GSA-run America.gov uses EO 14434's new term on its own pages. Its terms of use, under a dateline three weeks before the order, carry a section headed 'Answers generated with SI'. Its About page lists four press articles under link text reading 'SI-powered', 'SI Tool', 'SI-powered hub' and 'SI Push', while the four links' own web addresses read 'ai-powered', 'ai-tool', 'ai-powered-hub' and 'ai-push'; this desk did not open the articles and does not establish what their headlines say. Executive Order 14432, the America.gov order dated the same day as EO 14434, uses the term once, as lower-case 'super intelligence', without defining it or citing EO 14434, whose own definition is stated to apply 'For purposes of this order'. These are readings of one agency site and one order: they show the term in use, not which of EO 14434's two readings — Section 1's 'will not acknowledge' policy or Section 2's substitution in non-statutory documents — agencies follow in general.
Evidence limits. Readings of a live agency site and two printed orders. The terms page's dateline against its use of 'SI' is recorded as an open inconsistency, and nothing here shows that GSA altered anyone's headline — only what its page displays and where its links point.
Oct 2, 2026
Executive Order 14434 prints in the Federal Register at 91 FR 63129, word for word as filed for public inspection →records release · 2 source records
Read summary and limits
The Federal Register published Executive Order 14434, 'Inaugurating the Era of Super Intelligence', on October 2 at 91 FR 63129-63130, the date its public-inspection copy had scheduled. Compared word by word with that copy, the printed text is the same 687 words in the same order, by this desk's count, which ignores line breaks, page furniture, hyphenation and quotation-mark style; the Federal Register's API record lists no corrections. The terminology directive, the definition by cross-reference to 15 U.S.C. 9401(3) and the 60-day deadline for proposed legislative language, which runs from the order's September 29 date and which this desk counts to November 28, stand as filed.
Evidence limits. Publication of an order already on this file. The word comparison is this desk's, and printing changes nothing about what the order requires or establishes.
Oct 1, 2026
Executive Order 14434 tells agencies to write 'Super Intelligence' instead of 'AI' and orders a 60-day proposal for a new federal definition →records release · 2 source records
Read summary and limits
Executive Order 14434, 'Inaugurating the Era of Super Intelligence', dated at the White House September 29, went on Federal Register public inspection and is scheduled to print the next day. Section 2 directs agencies, to the maximum extent permitted by law, to use 'Super Intelligence' and 'SI' in place of 'Artificial Intelligence' and 'AI' in official correspondence, public communications, websites, reports, policy documents and other non-statutory documents, and exempts previously issued regulations, Presidential actions, contracts, grants and historical documents; the broader policy in Section 1 says the executive branch will not acknowledge the old terms' usage in any applicable setting. Section 3 defines the new terms by cross-reference, as the technologies already defined as artificial intelligence at 15 U.S.C. 9401(3), until superseded by later Presidential action or an Act of Congress. Read against that statute, enacted as section 5002 of division E of Public Law 116-283, the definition is structural and carries no capability threshold: a machine-based system that, for a given set of human-defined objectives, makes predictions, recommendations or decisions influencing real or virtual environments. The order gives the Assistant to the President for Science and Technology 60 days to send the President proposed legislative language for a federal definition reflecting the capabilities its Section 1 describes, with an assessment of whether it should modify, expand upon or supersede the statutory one. It sets no safety, standards, evaluation or procurement requirement, says it is subject to the availability of appropriations, and creates no enforceable right.
Evidence limits. A terminology directive and a request for draft legislation. It makes no finding about any AI system, company, agency or official, and an executive order cannot change the statutory definition it cites.

Read the full timeline →

Read the case scope: what is included and excluded

Included in this file

Executive-order and White House AI-governance records tracked by their current in-force or revoked status (EO 14110, EO 14148, EO 14179, America's AI Action Plan, and EO 14409).
NIST AI Risk Management Framework records and the Center for AI Standards and Innovation, formerly the U.S. AI Safety Institute, as federal AI standards and evaluation context.
Federal Trade Commission civil enforcement records involving alleged deceptive or unfair AI claims and their later dispositions.
Government Accountability Office oversight and audit reports on federal AI governance, implementation, and use.
Congressional AI oversight records: hearings and AI-governance legislation with sponsor, committee, and status detail.

Outside this file’s scope

Tracking AI model releases, vendor benchmark claims, or model capability announcements, and rating or ranking AI model quality, safety, or performance, which are covered by the AI release-receipts case.
Tracking federal AI procurement such as contract awards, OMB and FAR acquisition rules, USAspending, or SAM.gov records, which are covered by the federal AI procurement watch case.
Treating a policy, framework, hearing, or enforcement filing as proof that a specific AI system is safe, unsafe, compliant, or noncompliant.

What remains unresolved

These questions stay open in the case file.

Which specific EO 14110 provisions were rescinded versus continued by agency action after the EO 14179 review, and is there an agency-by-agency Federal Register trail?
CAISI's September 17, 2026 GLM-5.3 assessment documents the method behind its cyber benchmark comparisons (a ReAct agent harness, maximum reasoning settings, a one-parameter IRT capability index); what is its published methodology for other evaluation areas, and which developer agreements are active as of mid-2026?

Read all 12 open questions →

Connection map

How the public record connects

Graph JSON

Federal AI Safety GovernanceCase file

The White House and Executive Office of the Presidentagency Executive Order 14434 prints in the Federal Register at 91 FR 63129, word for word as filed for public inspectionOct 2, 2026 Executive Order 14434 tells agencies to write 'Super Intelligence' instead of 'AI' and orders a 60-day proposal for a new federal definitionOct 1, 2026 GovInfo (U.S. Government Publishing Office)public record repository EO 14179 Removing Barriers to American Leadership in AI signedJan 23, 2025 EO 14148 revokes EO 14110Jan 20, 2025 National Institute of Standards and Technologyagency Government Accountability Officeoversight body Center for AI Standards and Innovationoversight body

Subject
Person or role
Event

This view starts with the roles that have the most source-record links in this file. Explore Role Web and Timeline for the full set. A line represents a shared source record; it does not establish conduct or a direct relationship.

Publishing limits

These are the standards every anchor on this file must clear.

Minimum source tier

official record
oversight report
audit report
legislative record
civil enforcement record

Living people

For named officials, witnesses, or companies, publish only source-backed public roles and exact record relationships; do not infer intent, influence, or blame.

Victim protection

Do not publish private, nonpublic, or personal data; enforcement targets are alleged unless a record shows an adjudicated outcome.

Social leads

Social posts and coverage can start a governance or enforcement lead, but published copy requires official executive, standards, oversight, legislative, court, or enforcement records.

No guilt by proximity

A model, vendor, agency, or official appearing in the same governance lane is context only and does not prove a violation, a safety outcome, or responsibility for a policy.

Source records

Read the records behind this file. Search titles, publishers, summaries, and extracted facts. Each source keeps its reading limits and links to the events and roles it supports.

Search source recordsSource typePublisherSource order

Showing 1–10 of 30 sources · 30 in this file

official recordOct 2, 2026
America.gov About page (General Services Administration) — read 2026-10-02
U.S. General Services Administration (America.gov)
America.gov's About page, read October 2, 2026. It says the National Design Studio has worked with GSA on the site and that 'Phase 1 launched on September 29, 2026', the date of Executive Orders 14432 and 14434. Its Press section lists four outside articles under link text that uses 'SI' — 'SI-powered', 'SI Tool', 'SI-powered hub', 'SI Push' — and the four links go to politico.com, wsj.com, foxnews.com and bloomberg.com addresses whose own paths read 'ai-powered', 'ai-tool', 'ai-powered-hub' and 'ai-push'.
How to use this source
A reading of a live GSA page. The link text is GSA's; the articles are their publishers', and this desk did not open them, so it does not establish what their headlines say — only that the page shows 'SI' where each article's own web address shows 'ai'. That difference is recorded as the page presents it, not explained. Executive Order 14434 directs agencies to use 'SI' on websites and exempts previously issued regulations, Presidential actions, contracts, grants and other historical documents; the order does not say whether a third party's headline falls in any of those.
5 extracted facts
- It states: 'Phase 1 launched on September 29, 2026, introducing the simplest and fastest way to access information from the government. Phase 2 begins in 2027, bringing the ability to apply, enroll, and track progress directly in chat.'
- The Press link 'Trump signs executive order to launch SI-powered ‘America.gov’' points to https://www.politico.com/news/2026/09/29/trump-signs-executive-order-to-launch-ai-powered-america-gov-01096992.
- The Press link 'Trump Administration Launches SI Tool Aimed at Streamlining Government Online Services' points to https://www.wsj.com/politics/policy/trump-administration-launches-ai-tool-aimed-at-streamlining-government-online-services-4a2e5767.
- The Press link 'Trump unveils America.gov as SI-powered hub for passports, Medicare enrollment and more' points to https://www.foxnews.com/politics/trump-unveils-america-gov-ai-powered-hub-passports-medicare-enrollment.
- The Press link 'Trump Launches America.gov in SI Push for Federal Services' points to https://www.bloomberg.com/news/articles/2026-09-29/trump-launches-america-gov-in-ai-push-for-federal-services.
1 timeline event cites this source
- On the day EO 14434 prints, GSA's America.gov already writes 'SI', including in its link text for four outside articles whose web addresses read 'ai' →Oct 2, 2026 · records release review
Read original source ↗
official recordOct 2, 2026
America.gov terms of use (General Services Administration) — read 2026-10-02
U.S. General Services Administration (America.gov)
America.gov's terms of use, read October 2, 2026 under the dateline 'Last updated: September 8, 2026'. A section headed 'Answers generated with SI' says some answers are generated with SI and 'can be incomplete or wrong', and tells users to check the linked source before acting, especially on money, immigration status, benefits or a deadline. The terms say America.gov is not the agency of record and disclaim any warranty that the service is error-free.
How to use this source
GSA's own terms, not a measurement of how often answers are wrong. The page is inconsistent in a way recorded here and left open by this desk: it uses 'SI', the term Executive Order 14434 directed agencies to adopt in an order dated September 29, under a dateline three weeks earlier, while the same page promises to 'update the date at the top' whenever the terms change. Either the term was on the site before the order or the page changed without its date; this desk cannot tell which.
6 extracted facts
- The page reads: 'Last updated: September 8, 2026'.
- It states: 'America.gov is an official service of the United States government, operated by the General Services Administration.'
- Under the heading 'Answers generated with SI' it states: 'Some answers on America.gov are generated with SI. They can be incomplete or wrong, and they can be out of date the moment a program changes.'
- It states: 'Check the linked source before you act on an answer, especially for anything involving money, immigration status, benefits, or a deadline.'
- It states: 'America.gov is not the agency of record. It does not decide your eligibility, approve a benefit, or change a filing.'
- It states: 'We will post any change to these terms on this page and update the date at the top.'
1 timeline event cites this source
- On the day EO 14434 prints, GSA's America.gov already writes 'SI', including in its link text for four outside articles whose web addresses read 'ai' →Oct 2, 2026 · records release review
Read original source ↗
official recordOct 2, 2026
Executive Order 14432, Streamlining Access to Government Services Through America.gov (91 FR 63121, FR Doc. 2026-20319)
Executive Office of the President (Federal Register)
Executive Order 14432, dated at the White House September 29, 2026, filed for public inspection October 1 and printed October 2 at 91 FR 63121-63123. It directs the Administrator of General Services, in coordination with the National Design Studio and OMB, to establish and operate America.gov as the single point of entry for covered federal services online, with Login.gov as its authentication service, and makes it policy to 'ensure that super intelligence used in connection with America.gov is accurate, reliable, and transparent'. Covered services are public-facing federal services with more than 100,000 users in 12 months that can be accessed or applied for online, excluding IRS tax filing, Department of War services and elements of the Intelligence Community. OMB owes agencies an implementation memorandum within 90 days.
How to use this source
An executive order directing agencies, not a contract or award: it names no vendor, model, system, cost or contract vehicle for America.gov's super intelligence, sets no measure of accuracy, reliability or transparency, is subject to the availability of appropriations and creates no enforceable right. 'Super intelligence' appears once, in Section 2(e), in lower case; the order neither defines it nor cites Executive Order 14434 of the same date, whose definition of the capitalized 'Super Intelligence' and 'SI' is stated to apply 'For purposes of this order'. The order contains neither 'artificial intelligence' nor 'AI' (this desk's search of the Federal Register raw text). The 90-day due date is not printed; counted from September 29, this desk puts it at December 28, 2026.
11 extracted facts
- Section 1 states: 'This order therefore directs the establishment of America.gov as the unified digital front door to the Federal Government for every individual in the United States seeking Federal information or services.'
- Section 1 continues: 'It will be a secure, intelligent, and service-oriented point of entry through which an individual may sign in, communicate in plain language, receive accurate answers, and, where authorized and technically available, complete Government transactions without being required to navigate the websites of multiple agencies.'
- Section 2(e) makes it policy to 'ensure that super intelligence used in connection with America.gov is accurate, reliable, and transparent'.
- Section 2(f) makes it policy to 'preserve existing lawful methods by which Americans may obtain Federal services, including in-person, telephone, mail, and agency-specific digital services, so that America.gov provides a better option but does not become the only option.'
- Section 3 states: 'The term "covered services" means public-facing Federal services that serve more than 100,000 users in a 12-month period and can be accessed or applied for online. The term excludes tax filing services provided by the Internal Revenue Service and services provided by the Department of War and elements of the Intelligence Community.'
- Section 4 directs the Administrator of General Services, 'in coordination with the Administrator of the National Design Studio and the Director of OMB', to 'establish and operate America.gov as the single point of entry for Americans to access covered services online' and to integrate Login.gov with America.gov, 'including using Login.gov as the authentication service for America.gov.'
- Section 5(a) states: 'Within 90 days of the date of this order, the Director of OMB shall issue a memorandum to the heads of agencies on implementation of this order.'
- Section 5(b)(iv) directs heads of agencies to 'provide, on an ongoing basis, the Administrator of General Services and the Director of OMB with historical and current usage and performance data for covered services.'
- Section 6(b) states: 'This order shall be implemented consistent with applicable law and subject to the availability of appropriations.'
- The Federal Register raw text ends: '[FR Doc. 2026-20319 Filed 10-1-26; 11:15 am]'.
- The Federal Register API record for 2026-20319 gives executive_order_number '14432', citation '91 FR 63121', start_page 63121, end_page 63123, signing_date 2026-09-29 and publication_date 2026-10-02.
1 timeline event cites this source
- On the day EO 14434 prints, GSA's America.gov already writes 'SI', including in its link text for four outside articles whose web addresses read 'ai' →Oct 2, 2026 · records release review
2 roles are linked to this source
A shared source does not establish conduct or a direct personal relationship.
- The White House and Executive Office of the President →official record holder
- GovInfo (U.S. Government Publishing Office) →public record repository
Read original source ↗
official recordOct 2, 2026
Executive Order 14434, Inaugurating the Era of Super Intelligence, as printed (91 FR 63129, FR Doc. 2026-20321)
Executive Office of the President (Federal Register)
The printed Federal Register text of Executive Order 14434, published October 2, 2026 at 91 FR 63129-63130, the publication date its public-inspection copy had scheduled. This desk compared the printed PDF with the public-inspection PDF filed October 1 word by word and found the same 687 words in the same order.
How to use this source
The comparison and the word count are this desk's, not the Federal Register's. They ignore line breaks, page headers and footers, hyphens (so a word split across a line and a hyphenated word read alike) and the difference between straight and typographic quotation marks; within those limits the two texts are identical, and the Federal Register raw text differs from the PDF only by a placeholder where the signature graphic sits. The API record listed no corrections and no executive-order notes when read. Printing does not move the 60-day deadline, which runs from the order's September 29 date.
5 extracted facts
- The Federal Register API record for 2026-20321 gives executive_order_number '14434', citation '91 FR 63129', start_page 63129, end_page 63130, signing_date 2026-09-29, publication_date 2026-10-02 and an empty corrections list.
- The printed page header reads: 'Federal Register / Vol. 91, No. 190 / Friday, October 2, 2026 / Presidential Documents 63129'.
- Section 3(a) as printed reads: 'For purposes of this order, and except where otherwise provided by law, the terms "Super Intelligence" and "SI" mean the technologies and systems encompassed by the term "artificial intelligence" as defined in section 9401(3) of title 15, United States Code.'
- Section 3(b) as printed reads: 'Within 60 days of the date of this order, the Assistant to the President for Science and Technology (APST), in consultation with the heads of other agencies as the APST deems appropriate, shall submit to the President proposed legislative language to establish a Federal definition of "Super Intelligence" and "SI" that reflects the capabilities described in section 1 of this order.'
- The Federal Register raw text ends: 'THE WHITE HOUSE, September 29, 2026. [FR Doc. 2026-20321 Filed 10-1-26; 11:15 am]'.
2 timeline events cite this source
- Executive Order 14434 prints in the Federal Register at 91 FR 63129, word for word as filed for public inspection →Oct 2, 2026 · records release
- On the day EO 14434 prints, GSA's America.gov already writes 'SI', including in its link text for four outside articles whose web addresses read 'ai' →Oct 2, 2026 · records release review
2 roles are linked to this source
A shared source does not establish conduct or a direct personal relationship.
- The White House and Executive Office of the President →official record holder
- GovInfo (U.S. Government Publishing Office) →public record repository
Read original source ↗
official recordOct 1, 2026
Federal Register public inspection: Executive Order 14434, Inaugurating the Era of Super Intelligence (FR Doc. 2026-20321)
Executive Office of the President (Federal Register public-inspection copy)
The three-page public-inspection copy of Executive Order 14434, dated at the White House September 29, 2026, filed for public inspection October 1 at 11:15 a.m. Eastern and scheduled to print in the Federal Register October 2. It directs executive departments and agencies, to the maximum extent permitted by law, to use 'Super Intelligence' and 'SI' in place of 'Artificial Intelligence' and 'AI' in non-statutory documents; defines the new terms, for now, as the technologies already defined as artificial intelligence in 15 U.S.C. 9401(3); and gives the Assistant to the President for Science and Technology 60 days to submit proposed legislative language for a federal definition. It sets no AI safety, standards, evaluation or procurement requirement.
How to use this source
Public inspection is not publication, and the printed text could differ in pagination or formatting. The order's Section 1 policy statement reaches further than its Section 2 directive: the policy says the executive branch 'will not acknowledge the usage of' the old terms in any applicable setting, while Section 2 directs substitution only in non-statutory documents and Section 2(b) exempts previously issued regulations, Presidential actions, contracts, grants and historical documents. The 60-day due date is not printed in the order; counted from its September 29 date, this desk puts it at November 28, 2026. The statutory definition the order points at is anchored separately below. An executive order cannot amend a statute, so the statutory definition of 'artificial intelligence' stands unless Congress acts.
9 extracted facts
- Section 1 states the policy that 'to the maximum extent permitted by law, the executive branch shall use the terms "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI" and will not acknowledge the usage of "Artificial Intelligence" and "AI" in any applicable setting.'
- Section 2(a) directs that agencies 'shall use "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI" in official correspondence, public communications, websites, reports, policy documents, and other non-statutory documents within the executive branch.'
- Section 2(b) states: 'Nothing in this section requires the alteration of previously issued regulations, Presidential actions, contracts, grants, or other historical documents.'
- Section 3(a) states the terms 'mean the technologies and systems encompassed by the term "artificial intelligence" as defined in section 9401(3) of title 15, United States Code', governing 'unless and until superseded by subsequent Presidential action consistent with applicable law or by an Act of Congress.'
- Section 3(b) states: 'Within 60 days of the date of this order, the Assistant to the President for Science and Technology (APST), in consultation with the heads of other agencies as the APST deems appropriate, shall submit to the President proposed legislative language to establish a Federal definition of "Super Intelligence" and "SI" that reflects the capabilities described in section 1 of this order.'
- Section 4 states: 'This order shall be implemented consistent with applicable law and subject to the availability of appropriations.', that it does not 'create any right or benefit, substantive or procedural, enforceable at law or in equity by any party', and that 'The costs for publication of this order shall be borne by the Office of Science and Technology Policy.'
- The filed PDF ends: '[FR Doc. 2026-20321 Filed: 10/1/2026 11:15 am; Publication Date: 10/2/2026]'.
- The Federal Register's public-inspection API record for 2026-20321 gives filed_at 2026-10-01T11:15:00.000-04:00, filing_type 'special', type 'Presidential Document', num_pages 3, publication_date 2026-10-02, agency 'Executive Office of the President', and no docket numbers.
- The order names no company, model, agency program or AI system, and states no safety, standards, evaluation, reporting or procurement requirement.
2 timeline events cite this source
- Executive Order 14434 prints in the Federal Register at 91 FR 63129, word for word as filed for public inspection →Oct 2, 2026 · records release
- Executive Order 14434 tells agencies to write 'Super Intelligence' instead of 'AI' and orders a 60-day proposal for a new federal definition →Oct 1, 2026 · records release
1 role is linked to this source
A shared source does not establish conduct or a direct personal relationship.
- The White House and Executive Office of the President →official record holder
Read original source ↗
audit reportSep 24, 2026
Assessment of Artificial Intelligence Risk Management Practices (TIGTA Report No. 2026-208-059)
Treasury Inspector General for Tax Administration
TIGTA final audit report dated September 24, 2026 on how the IRS documents risk management and data quality for AI use cases it presumed high-impact under OMB Memorandum M-25-21, whose minimum practices were due by April 2026. Reviewing five deployed use cases, it found two without a documented impact assessment and four without testing documentation, and made two recommendations, both agreed to by the IRS.
How to use this source
An audit of documentation for five use cases, not of whether any IRS AI system produced wrong results; TIGTA says it did not evaluate the accuracy of the data itself, and IRS analytics staff told it data-quality checks were completed. The statement that Treasury had issued no implementing guidance is attributed to IRS management.
5 extracted facts
- It states: 'As of December 2025, the IRS had an inventory of 225 AI use cases.'
- It states: 'Our review of 5 deployed high-impact AI use cases found that 40 percent (2 of 5) AI use cases did not have an impact assessment documented.'
- On data-quality testing it states: 'The remaining 80 percent (4 of 5) did not have any testing documentation.'
- It states: 'RAAS management stated that as of July 2026, they have not received any guidance or instructions from the Department of the Treasury to implement minimum risk management practices.'
- It states: 'The IRS agreed with both recommendations and stated that it either has or plans to implement corrective actions.'
1 timeline event cites this source
- TIGTA finds the IRS lacked impact assessments for two of five high-impact AI uses it reviewed under OMB M-25-21 →Sep 24, 2026 · audit report
1 role is linked to this source
A shared source does not establish conduct or a direct personal relationship.
- Treasury Inspector General for Tax Administration →audit authority
Read original source ↗
legislative recordSep 23, 2026
Senate Judiciary Subcommittee on Crime and Counterterrorism hearing: Always Watching: Flock's Nationwide AI Surveillance Network
U.S. Senate Committee on the Judiciary
The committee's hearing page for a subcommittee hearing held Wednesday, September 23, 2026 at 2:30 p.m. in Dirksen Senate Office Building Room 562. Five witnesses are listed: legislative counsel for the Institute for Justice, the Sheriff of Pinal County, Arizona, a cybersecurity engineer, a senior policy counsel from the ACLU, and one private individual.
How to use this source
A hearing builds the oversight record. Witness testimony gives the witnesses' views and is not a finding by Congress or any agency. Nothing here establishes that Flock or any law-enforcement user acted unlawfully. The private witness is neither named nor characterized here under this desk's victim-protection rule.
3 extracted facts
- The page lists the hearing type as 'Subcommittee Hearing' of the Subcommittee on Crime and Counterterrorism.
- The page gives the date as 'Wednesday, September 23rd, 2026', time 02:30pm, location Dirksen Senate Office Building Room 562.
- The page lists five witnesses with downloadable testimony, including the Pinal County, Arizona Sheriff and an ACLU senior policy counsel.
1 timeline event cites this source
- Senate Judiciary subcommittee holds a hearing on Flock's AI surveillance network →Sep 23, 2026 · records release
1 role is linked to this source
A shared source does not establish conduct or a direct personal relationship.
- Senate Judiciary Subcommittee on Crime and Counterterrorism →legislative oversight authority
Read original source ↗
official recordSep 17, 2026
CAISI's Assessment of Z.ai's GLM-5.3 Cyber Capabilities
National Institute of Standards and Technology / Center for AI Standards and Innovation
CAISI's September 17, 2026 update assessing the cyber capabilities of GLM-5.3, a model the PRC-based company Z.ai released August 14, 2026 and whose weights it published two weeks later. CAISI evaluated GLM-5.3 on four benchmarks: SEC-Bench Pro, ExploitBench, ExploitGym (Userspace) and the private CAISI OSS-Fuzz. It compared the results with the best U.S. and PRC scores it has recorded and reported an aggregate 'cyber capability index' fitted with a one-parameter Item Response Theory model.
How to use this source
This is a government evaluation of one model on CAISI's chosen benchmarks. It is not a safety certification, a ban, or a finding about any company's conduct. The 'U.S. frontier best' comparison takes the best score of any U.S. model CAISI has evaluated, including trusted-access releases, and U.S. models were 'when applicable' tested with cyber safeguards disabled, so the comparison does not reflect the public products people use. Per this file's rules, it records what CAISI measured and stated, not a rating by this desk.
7 extracted facts
- CAISI states 'GLM-5.3 is the most cyber-capable open-weight model released to date.'
- CAISI states GLM-5.3 'lags the capability level of the U.S. frontier by about four months in an aggregate measure of performance across CAISI cyber benchmarks.'
- SEC-Bench Pro: GLM-5.3 40.4% (74/183), U.S. frontier best 90.2% (165/183), PRC frontier best 27.3% (50/183).
- ExploitBench: GLM-5.3 61.1% (9.8/16), U.S. frontier best 100.0% (16.0/16), PRC frontier best 32.2% (5.1/16).
- ExploitGym (Userspace): GLM-5.3 9.4% (47/498), U.S. frontier best 44.4% (223/502), PRC frontier best 2.6% (13/502); OSS-Fuzz: 7.7% (23/297), 23.2% (69/297), 2.4% (7/297).
- CAISI states models were evaluated 'as agents in a ReAct harness with bash, python, and a nudge to continue if stopped,' at maximum reasoning settings, and 'when applicable, U.S. models were tested with cyber safeguards disabled.'
- CAISI states a 400-point increase in its cyber capability index 'equates to a 10x increase in the statistical odds of solving tasks.'
1 timeline event cites this source
- CAISI publishes a cyber-capability assessment of Z.ai's open-weight GLM-5.3 →Sep 17, 2026 · records release
1 role is linked to this source
A shared source does not establish conduct or a direct personal relationship.
- Center for AI Standards and Innovation →official record holder
Read original source ↗
civil enforcement recordAug 27, 2026
FTC finalizes orders with Cox Media Group and two other firms over 'Active Listening' AI-powered marketing claims
Federal Trade Commission
FTC release of August 27, 2026: 'The Federal Trade Commission finalized orders requiring Cox Media Group (CMG) and two other firms to pay a total of $930,000 to settle allegations' that they deceived customers about an 'Active Listening' AI-powered marketing service.
How to use this source
Consent orders settle allegations; they are not adjudicated findings that the conduct occurred.
2 extracted facts
- CMG must pay $880,000 while both MindSift and 1010 Digital Works must each pay $25,000.
- The Commission voted 2-0 to give final approval to the consent agreements.
1 timeline event cites this source
- FTC finalizes 'Active Listening' AI marketing consent orders →Aug 27, 2026 · civil settlement
Read original source ↗
official recordJul 1, 2026
FTC seeks public comment on a proposed policy statement addressing AI accuracy
Federal Trade Commission
FTC release of July 1, 2026 seeking comment on a proposed policy statement stating that 'AI companies that distort their systems' outputs to achieve undisclosed ideological objectives could be deceiving consumers in violation Section 5 of the FTC Act.'
How to use this source
A proposed policy statement open for comment. It is not an enforcement action, a rule, or a finding against any company, and its preemption view is the Commission's stated position rather than a court holding.
4 extracted facts
- The release is dated July 1, 2026.
- The Commission vote authorizing the Federal Register notice was 2-0.
- The public had until July 31, 2026 to submit comments on the policy statement.
- The release states that the proposed policy statement explains a state law of the kind it addresses is 'impliedly preempted to the extent it conflicts with a federal regulatory scheme.'
1 timeline event cites this source
- FTC seeks comment on a proposed AI-accuracy policy statement →Jul 1, 2026 · records release
Read original source ↗

20 more in this view

Import atlas

Public sources ready to ingest

Source atlas

sources27

keyless23

matched anchors0

credential gated2

ready keylessEleventh Circuit Epstein CVRA opinions0 matched case anchorsRun ca11-epstein-cvra-opinions-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessSecond Circuit Maxwell appeal opinions0 matched case anchorsRun ca2-maxwell-appeal-opinions-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessCongress.gov Epstein Files Transparency Act bill record0 matched case anchorsRun congress-gov-epstein-files-transparency-act-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessDOJ Epstein EFTA Congress reports0 matched case anchorsRun doj-epstein-congress-reports-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessDOJ Epstein official court records index0 matched case anchorsRun doj-epstein-court-records-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessDOJ Epstein Library0 matched case anchorsRun doj-epstein-library-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessDOJ Epstein EFTA review protocols0 matched case anchorsRun doj-epstein-review-protocols-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

ready keylessDOJ OIG Epstein oversight records0 matched case anchorsRun doj-oig-epstein-oversight-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

Import sources create source-run packets, document inventories, and review tokens. They do not publish accusations or graph relationships until the case-file evidence rules are satisfied.

Source validation

Review snapshot

Report Markdown

Case baselinebaseline

known30

publishable28

queued0

candidates0

quarantine0

Deduped0Submitted URLs already in the case-file anchors.

Blocked0Invalid or non-public URLs rejected from review.

No community submissions are in the current review snapshot.

official record reviewOfficial records can become source anchors after exact-record review.

court docket reviewCourt records need docket, filing, order, or document metadata attached.

legislative release reviewLegislative releases need transcript, subpoena, letter, or hearing context.

lead quarantineLeads stay out of timelines and graphs until joined to public records.

Filed: Hugin · public social evidence reports
License: Text and data are shared under the terms noted at hugin.studio/support · attribution appreciated.
Fingerprint: A583 F601 EBF3

For researchers & contributorsSource desk, research joins & machine-readable exportsEverything Hugin uses to grow and audit this file — connection detail, review posture, scout queue, and every export.

Data exports

Every machine-readable slice of this file

Connection map

How the public record is joined

Graph JSON

timeline backed24

source links32

role links31

open queues2

Source groups

17 sourcesofficial record17 linked recordsNIST AI Risk Management Framework (AI RMF 1.0) / NIST AI RMF Generative AI Profile (NIST AI 600-1) / Center for AI Standards and Innovation (CAISI)

4 sourcescivil enforcement record4 linked recordsFTC Operation AI Comply enforcement sweep / Rytr LLC; Analysis of Proposed Consent Order To Aid Public Comment (89 FR 80565) / FTC reopens and sets aside the Rytr final order

Timeline joins

Jan 1, 2021 / officialCongress enacts the federal definition of 'artificial intelligence' at 15 U.S.C. 9401(3), with no capability threshold in it1 linked recordA statutory definition read in the 2024 Code edition, which this page states is current through January 6, 2025; it is not a finding about any system and this desk confirmed no later amendment independently.

Jan 26, 2023 / officialNIST releases the AI Risk Management Framework1 linked recordA voluntary framework does not establish whether any agency or vendor met it.

Role joins

official record holderNational Institute of Standards and Technology4 linked recordsNIST guidance is voluntary and non-regulatory; publishing a framework is not the same as mandating or certifying compliance.

official record holderCenter for AI Standards and Innovation2 linked recordsReconstituted from the U.S. AI Safety Institute in June 2025 with a reframed mission; continuity of specific programs across the rename should be verified per program.

Connections are navigation aids. They show where a record, event, or role is sourced; they do not convert association, correspondence, testimony requests, or document appearance into misconduct findings.

Source desk

Scout queue before promotion

Promotion readiness Source scout JSON Scout summary Scout CSV Scout JSONL Scout Markdown Connector matrix Matrix Markdown Packet JSON Packet Markdown

verified anchors30

timeline events24

research queues7

No hand-tuned source desk on this case yet, so no lead or probe queue is counted for it.

scout targets16

first-pass scout14

credential queue2

connectors16

keyless APIs6

free-key targets2

public search8

official pages0

POST templates0

CourtListener Search API / keyless apiGDELT DOC API / keyless apiFederal Register API / keyless apiGovInfo public document search / public searchRegulations.gov public search / public searcheCFR public search / public searchDOJ News API / keyless apiFOIA.gov government website search / public searchFBI Vault search / public searchNational Archives Catalog API / free key requiredData.gov CKAN catalog API / free key requiredCongress.gov public search / public search

Default public-record source desk for cases without a hand-tuned scout profile.

Source lanespublic-record-discovery

Sharing stateautomatic sharing disabled

Next actionAdd source receipts to the review queue before promoting case claims or social copy.

keyless apiCourtListener Search APIpublic-record source deltasLegal search across opinions, dockets, RECAP data, judges, and arguments.

keyless apiGDELT DOC APIpublic-record source deltasCoverage-index discovery for fresh public reporting rows.

keyless apiFederal Register APIpublic-record source deltasOfficial rule, notice, and agency-document discovery.

public searchGovInfo public document searchpublic-record source deltasFree public search across official publications from all three branches of the U.S. federal government.

public-record source deltas

Source leads and automated probes are review inventory. They become verified anchors only after Hugin can attach the public record, legal posture, and no-allegation guardrails.

Join board

Public-record joins are wired into exports

These are not loose TODOs anymore. Each lane below is computed from the case file's source tiers, then joined to matching sources, timeline events, role nodes, and release-roadmap requests.

Source atlas Review queue Release board

wired lanes5

review lanes2

source joins49

release joins0

Wired lanes5Computed from the source tiers and joined to live records, roles, and releases.

WiredExecutive-order and AI Action Plan status monitor

18 sources

Track new or amended AI executive orders and AI Action Plan implementation through GovInfo and White House records before saying federal AI policy changed.

timeline15

roles6

release asks0

Matched records

NIST AI Risk Management Framework (AI RMF 1.0)official record / National Institute of Standards and Technology NIST AI RMF Generative AI Profile (NIST AI 600-1)official record / National Institute of Standards and Technology Center for AI Standards and Innovation (CAISI)official record / National Institute of Standards and Technology Executive Order 14110 on Safe, Secure, and Trustworthy AIofficial record / The White House (GovInfo Compilation of Presidential Documents)

Connected roles

National Institute of Standards and TechnologyCenter for AI Standards and InnovationThe White House and Executive Office of the President

Release requests

No release request is attached to this lane yet.

PublishesDated in-force or revoked status of each order with GovInfo or Federal Register citations.

BlockedPredicting future policy or characterizing an administration's intent.

Graph edges Source checks

WiredCAISI evaluation and agreement tracking

17 sources

Monitor NIST and CAISI records for evaluation reports, developer agreements, and standards initiatives.

timeline15

roles6

release asks0

Matched records

Connected roles

National Institute of Standards and TechnologyCenter for AI Standards and InnovationThe White House and Executive Office of the President

Release requests

No release request is attached to this lane yet.

PublishesNamed, dated CAISI actions with NIST citations.

BlockedRating model quality or safety, or implying a pass or fail verdict from the existence of an evaluation.

Graph edges Source checks

WiredFTC AI enforcement lifecycle

5 sources

Track FTC AI matters end to end using Federal Register consent notices and regulations.gov dockets, recording both openings and later set-asides.

timeline3

roles3

release asks0

Matched records

GovInfo catalog record for NIST AI 100-1 (AI RMF 1.0)public record repository / U.S. Government Publishing Office (GovInfo)FTC Operation AI Comply enforcement sweepcivil enforcement record / Federal Trade Commission Rytr LLC; Analysis of Proposed Consent Order To Aid Public Comment (89 FR 80565)civil enforcement record / Federal Trade Commission / U.S. Government Publishing Office FTC reopens and sets aside the Rytr final ordercivil enforcement record / Federal Trade Commission

Connected roles

National Institute of Standards and TechnologyFederal Trade CommissionGovInfo (U.S. Government Publishing Office)

Release requests

No release request is attached to this lane yet.

PublishesDocket-numbered actions, parties, dates, and current status framed as allegations or orders.

BlockedStating a company is guilty, or omitting a later reversal such as the Rytr set-aside.

Graph edges Source checks

WiredGAO AI oversight and recommendation tracking

4 sources

Monitor GAO's AI topic for new reports and open or closed recommendation status.

timeline3

roles2

release asks0

Matched records

GAO AI Accountability Framework (GAO-21-519SP)audit report / Government Accountability Office GAO report on AI management and personnel requirements (GAO-24-107332)oversight report / Government Accountability Office GAO generative AI use and management report (GAO-25-107653)oversight report / Government Accountability Office Assessment of Artificial Intelligence Risk Management Practices (TIGTA Report No. 2026-208-059)audit report / Treasury Inspector General for Tax Administration

Connected roles

Government Accountability OfficeTreasury Inspector General for Tax Administration

Release requests

No release request is attached to this lane yet.

PublishesReport-numbered findings and recommendation status with GAO citations.

BlockedTracking procurement spending or contract awards, which belong to the federal AI procurement case.

Graph edges Source checks

WiredFederal AI-governance legislation and hearings

3 sources

Monitor congress.gov and GovInfo for AI-governance bills and oversight hearings across the 118th and 119th Congresses.

timeline3

roles4

release asks0

Matched records

Senate hearing: Oversight of A.I. Rules for Artificial Intelligencelegislative record / U.S. Senate Judiciary Subcommittee on Privacy, Technology, and the Law S. 3312 Artificial Intelligence Research, Innovation, and Accountability Act of 2024legislative record / U.S. Congress (GovInfo BILLS)Senate Judiciary Subcommittee on Crime and Counterterrorism hearing: Always Watching: Flock's Nationwide AI Surveillance Networklegislative record / U.S. Senate Committee on the Judiciary

Connected roles

Senate Judiciary Subcommittee on Privacy, Technology, and the LawSenate Judiciary Subcommittee on Crime and CounterterrorismSenate Committee on Commerce, Science, and Transportation

Release requests

No release request is attached to this lane yet.

PublishesBill numbers, sponsors, committee action, enacted status, and hearing records.

BlockedPredicting passage or inferring a lawmaker's motive.

Graph edges Source checks

Needs review2Sources have landed but the lane is not promoted to wired yet.

Needs reviewFederal Register and GovInfo AI-document index

1 source

Build a citation-anchored index of AI executive orders, agency notices, and requests for information through GovInfo, since federalregister.gov blocks automated fetch.

timeline0

roles2

release asks0

Matched records

GovInfo catalog record for NIST AI 100-1 (AI RMF 1.0)public record repository / U.S. Government Publishing Office (GovInfo)

Connected roles

National Institute of Standards and TechnologyGovInfo (U.S. Government Publishing Office)

Release requests

No release request is attached to this lane yet.

PublishesA document list with Federal Register or GovInfo citations and dates.

BlockedSummaries not traceable to a fetched primary record.

Graph edges Source checks

Needs reviewInternational index cross-check (OECD.AI)

1 source

Reconcile OECD.AI United States entries against U.S. primary records, including the observed EO 14179 signing-date discrepancy.

timeline0

roles1

release asks0

Matched records

OECD.AI Policy Observatory entry for Executive Order 14179coverage index / OECD.AI Policy Observatory

Connected roles

OECD.AI Policy Observatory

Release requests

No release request is attached to this lane yet.

PublishesFlagged discrepancies with the governing primary citation.

BlockedTreating the index as a primary source.

Graph edges Source checks

Public utility

Help grow the source queue

Hugin accepts public records, not accusations. A submitted URL lands in a review lane, gets source-validated, and only becomes a case anchor after the evidence rules and privacy gates are satisfied.

Submit source Public queue

Review posture

What this case can prove right now

Fact-check desk

publishable claims24

source repair0

known anchors30

lead quarantine0

scout targets16

first-pass scout14

credential queue2

scout connectors16

rankmedium124 priority score

readyReady for first-pass public review24 publishable claim rows / 30 source anchors / 14 first-pass source rows

actions6 start-here rows9 total workbench action rows

blockers1 launch checks2 source-scout rows require free operator keys

start-here / hash-bundle-firstOpen verification packet MarkdownCombines claims, source anchors, first-pass rows, and hashes before any launch copy is written.start-here / claim-text-before-copyReview claim ledger MarkdownShows which timeline-derived claims are publishable and which still need source repair.start-here / claim-to-source-before-shareAudit claim-source matrix MarkdownJoins each claim row to its exact source anchor, source tier, publisher, and review warning before quoting.next / anchor-row-before-quoteReview source anchors MarkdownShows the flat public-record anchor rows, publishers, extracted facts, and timeline links used by the case.next / lead-before-anchorRun first-pass source-scout rowsSurfaces credentialless/keyless review rows before credential and POST-template queues.next / connector-family-before-scale-claimAudit source-scout connector matrixShows connector-family coverage, access classes, case coverage, and credential gates before citing source-desk scale.

review feedCase workbench JSONRanked launch-readiness actions for this case.review feedCase workbench MarkdownReadable start-here handoff for case review.review feedClaim ledger JSONEvidence-status rows derived from the case timeline.review feedClaim ledger MarkdownReadable review queue for source-backed claims.review feedSource validationPromotion, dedupe, and quarantine checks.review feedVerification packet JSONComplete case-scoped packet for source review and fact checks.review feedVerification packet MarkdownReadable handoff for public review and launch notes.review feedSource scout summaryFirst-pass, credential, and manual POST queue split.review feedSource scout CSVConnector targets still waiting for review.review feedSource scout JSONLLine-delimited scout rows for audits and imports.review feedSource scout MarkdownReadable source-scout handoff for manual review.review feedConnector matrix JSONCase-scoped connector-family coverage, access classes, and credential gates.review feedConnector matrix MarkdownReadable connector-family audit before citing source-desk scale.review feedFact-check deskGlobal civic review posture and verification URLs.

Fact-check status is a review posture, not a verdict. Source-scout rows are leads until a reviewer joins them to publishable anchors; claim-ledger rows are timeline-derived and still require source validation plus no-allegation guardrails before promotion.

Source intake

Queue a public record

Public queue

Paste an official or public-record URL. Hugin classifies the lane first, keeps lead-only material quarantined, and requires source validation before anything reaches the case file.

Queue a public source for review.

Visible queue

Latest submitted public URLs

Sanitized queue view: Hugin shows the public URL, lane, status, and why a source has not been promoted. Submitter identifiers are not shown.

Queue JSON Markdown Change ledger

Public queue loaded.Live queue connectedQueue hash c4f3c8718a38

No public submissions in this snapshot.When someone submits a source, this board will show the URL, lane, status, hold reason, and next review step before anything can become part of the case file.

known sources30

publishable anchors28

atlas sources27

keyless feeds23

release asks7

document jobs30

privacy triage5

lead quarantine0

01Submit URLPublic URL, label, and context enter the source queue.

02Classify laneOfficial, court, legislative, civil, state, repository, or lead quarantine.

03Validate sourcePublisher, date, scope, privacy, and duplicate checks run before promotion.

04Promote anchorOnly source-backed, privacy-cleared records can join timeline or role surfaces.

Contribution guide

Submit sources, not claims

A source is a public URL someone else can inspect. A claim is an interpretation. Hugin queues sources first and keeps claims out until a reviewed public record supports the exact wording.

Accepted

Official agency pages, court dockets, congressional releases, civil records, state updates, and public repositories with source URLs.

Publisher, date, record family, and why the URL belongs in this file.

Source URLs first. Context is welcome, but claims do not publish from submissions alone.

Not accepted

Private tips, survivor identifiers, sealed or leaked material, graphic details, and private contact data.

Screenshots or social posts without a public source URL.

Misconduct labels for living people unless the exact cited public record supports that exact statement.

What changed

Case change ledger

Audit feed for source anchors, queue decisions, privacy-review jobs, and hashes. It explains the process without turning leads into claims.

Ledger JSON Markdown

Change ledger loaded.Ledger hash a9014a0ff2aeCase hash a583f601ebf3

Oct 21, 2026document privacy review

GovInfo catalog record for NIST AI 100-1 (AI RMF 1.0)

Status: needs human privacy review
Evidence state: metadata only
Lane: privacy triage

Extract document index and source metadata first.

Next: Metadata extraction includes publisher, URL, document family, and date.cc69257596cb

Oct 21, 2026document privacy review

OECD.AI Policy Observatory entry for Executive Order 14179

Status: needs human privacy review
Evidence state: metadata only
Lane: privacy triage

Extract document index and source metadata first.

Next: Metadata extraction includes publisher, URL, document family, and date.c7af906d855f

Oct 21, 2026document privacy review

S. 3312 Artificial Intelligence Research, Innovation, and Accountability Act of 2024

Status: needs human privacy review
Evidence state: metadata only
Lane: privacy triage

Extract document index and source metadata first.

Next: Metadata extraction includes publisher, URL, document family, and date.55a8c100ddb0

Oct 21, 2026document privacy review

Senate Judiciary Subcommittee on Crime and Counterterrorism hearing: Always Watching: Flock's Nationwide AI Surveillance Network

Status: needs human privacy review
Evidence state: metadata only
Lane: privacy triage

Extract document index and source metadata first.

Next: Metadata extraction includes publisher, URL, document family, and date.314ebf317ff0

Oct 21, 2026document privacy review

Senate hearing: Oversight of A.I. Rules for Artificial Intelligence

Status: needs human privacy review
Evidence state: metadata only
Lane: privacy triage

Extract document index and source metadata first.

Next: Metadata extraction includes publisher, URL, document family, and date.25a3c014e5b7

Oct 21, 2026hash snapshot

Case file hash snapshot

Status: hash changed when case bundle changes
Evidence state: metadata only

Mirrors can compare this hash to verify the same case file bundle.

Next: Compare the short hash in the UI with this machine-readable ledger.a583f601ebf3

Oct 21, 2026hash snapshot

Document manifest hash snapshot

Status: hash changed when document handling changes
Evidence state: metadata only

This binds document mirroring and privacy-triage rules.

Next: Compare the short hash in the UI with this machine-readable ledger.e147585b727c

Oct 21, 2026hash snapshot

Document review queue hash snapshot

Status: hash changed when document jobs change
Evidence state: metadata only

This binds privacy-review and metadata extraction jobs.

Next: Compare the short hash in the UI with this machine-readable ledger.2311316c60c8

Oct 21, 2026hash snapshot

Source validation hash snapshot

Status: hash changed when queue or validation rules change
Evidence state: metadata only

This binds the visible queue decisions to a machine-readable report.

Next: Compare the short hash in the UI with this machine-readable ledger.fbb6cf71c948

Oct 21, 2026hash snapshot

Submission queue hash snapshot

Status: hash changed when public queue rows change
Evidence state: metadata only

This changes when sanitized submission rows change.

Next: Compare the short hash in the UI with this machine-readable ledger.c4f3c8718a38

Safety posture

Locked to public records and bounded intake

Post throttleSource intake is DB-rate-limited to 6 submissions per minute per network key before a record is written.

Public URL onlyLocalhost, private networks, URL credentials, unsupported protocols, malformed URLs, and overlong fields are rejected.

No paid callsThe data export links render deterministic public-record payloads. They do not call LLMs or paid upstream APIs.

Identity minimizedSubmitter material is converted to a server-side HMAC hash, and the public queue does not return raw IPs or user agents.

Official-source wiring

Ready lanes to cook down

Full atlas

ready keylessEleventh Circuit Epstein CVRA opinions0 matched anchors / keylessRun ca11-epstein-cvra-opinions-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.ready keylessSecond Circuit Maxwell appeal opinions0 matched anchors / keylessRun ca2-maxwell-appeal-opinions-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.ready keylessCongress.gov Epstein Files Transparency Act bill record0 matched anchors / keylessRun congress-gov-epstein-files-transparency-act-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.ready keylessDOJ Epstein EFTA Congress reports0 matched anchors / keylessRun doj-epstein-congress-reports-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.ready keylessDOJ Epstein official court records index0 matched anchors / keylessRun doj-epstein-court-records-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.ready keylessDOJ Epstein Library0 matched anchors / keylessRun doj-epstein-library-import, store a source-run packet, and reconcile 0 matched case anchors before adding derived timeline rows.

Graph JSONMachine provenance nodes and edges, not a visual page.Submission queueMutable review queue for public source leads.Validation reportPromotion rules, dedupe checks, and lead quarantine.Verification packetCase hashes, claim gates, source anchors, and scout rows in one bundle.Change ledgerWhat changed, what stayed held, and which hashes prove it.Source atlasOfficial feeds, import lanes, and credential gates.Connector matrixCase-scoped connector families, access classes, and credential gates.Document manifestPrivacy and mirroring rules for source documents.Review queueMetadata, privacy, and safe-excerpt worker jobs.Mirror seedPublic node seed for independent source mirrors.

Start reading

Recent developments in the record

Included in this file

Outside this file’s scope

What remains unresolved

How the public record connects

Publishing limits

No guilt by proximity

People, institutions, and the records between them

Records in view 22

People 1

Courts & prosecutors 4

Oversight & legislature 6

Records & law 2

Timeline

Source records

Public sources ready to ingest

Review snapshot

What the record currently supports

What should be released next

What stays unresolved

Every machine-readable slice of this file

How the public record is joined

Scout queue before promotion

Public-record joins are wired into exports

Help grow the source queue

What this case can prove right now

Queue a public record

Latest submitted public URLs

Submit sources, not claims

Case change ledger

GovInfo catalog record for NIST AI 100-1 (AI RMF 1.0)

OECD.AI Policy Observatory entry for Executive Order 14179

S. 3312 Artificial Intelligence Research, Innovation, and Accountability Act of 2024

Senate Judiciary Subcommittee on Crime and Counterterrorism hearing: Always Watching: Flock's Nationwide AI Surveillance Network

Senate hearing: Oversight of A.I. Rules for Artificial Intelligence

Case file hash snapshot

Document manifest hash snapshot

Document review queue hash snapshot

Source validation hash snapshot

Submission queue hash snapshot

Locked to public records and bounded intake

Ready lanes to cook down