SOURCES / METHOD / LIMITS
Know what the
evidence supports.
AI Governance Radar helps people move from public source material to a reviewable next step. Each stage has a different meaning: a retrieved page, a suggested interpretation and an assessed control are separate kinds of evidence.
01 / SOURCE SELECTION
A selected catalog with visible origins.
The catalog contains 20 selected primary sources from regulators, public institutions, standards bodies and framework publishers. Records identify the publisher, jurisdiction, source type, original URL and editorial review date. Coverage is selective and is not a complete register of every AI obligation.
We distinguish legislation, public-sector policy, standards, voluntary guidance and initiatives. Inclusion does not mean a source applies to your organization. Public summaries of licensed standards do not replace access to the full standard.
The catalog’s editorial review date is . This methodology review does not refresh that date or claim that every source has been checked today. Open the original publication to confirm its current text and status.
02 / INTERPRET THE SIGNAL
A page change starts a review.
- A successful retrieval
- The monitor could obtain supported content from the source at a recorded time. The timestamp is evidence of that check, not of uninterrupted monitoring or complete coverage.
- A baseline or detected change
- A baseline establishes a comparison point. Supported HTML and text pages are compared using normalized text; PDFs are compared using document bytes, without extracting their text. A different fingerprint may reflect navigation, formatting, metadata or editorial changes. It does not establish a meaningful content change or a legal amendment.
- An unchanged source or failed retrieval
- Unchanged means the normalized text or PDF bytes match the prior successful comparison. A failure leaves current source status uncertain. Neither outcome proves that no relevant development occurred elsewhere.
- A summary or suggested action
- Curated developments and AI briefings help a person read the source. A quotation match can confirm that text appeared in retrieved material; it does not validate the interpretation or establish applicability.
Use the recorded retrieval times and failure information to judge freshness. A configured schedule is not proof that a check completed. Review the linked source before acting on a proposed obligation or deadline.
03 / ASSESS THE ORGANIZATION
Keep recorded claims and tested evidence distinct.
System matches are relevance filters using recorded context such as jurisdiction, role and sector. They help prioritize review; they do not decide legal scope. Control mappings are prompts for investigation, not equivalence certificates between frameworks.
Workspace metrics separate controls assessed, implementation reported and current evidence recorded. Evidence currency follows the dates and review information entered by the user. The application does not independently inspect the underlying evidence or certify the organization.
For a consequential decision, ask an appropriately qualified reviewer to check the source, system context, control design and test results. Document unresolved questions and who accepts the remaining risk.
The NIST AI RMF and its Playbook provide voluntary risk-management guidance. Our practical governance guide explains how to begin a review.
PRODUCT SCOPE
Public intelligence. A workspace on your device.
Public reference pages and the showcase contain public material, not your saved systems, assessments or tasks. Workspace records are stored in this browser. Manual exports are not encrypted; handle them according to your organization’s information policy.
The Activity view is an editable, bounded local history, not a tamper-evident audit ledger. Account signup, shared tenant storage, SSO, connected runtime monitoring and independent assurance are not implemented. Newsletter preferences do not create an organization account.
PRIMARY REFERENCES
Go to the original publication.
The links below reproduce the current catalog’s source list. Publisher content and availability can change after editorial review.
- Global · GuidanceOWASP GenAI LLM Top 10 2026OWASP GenAI Security Project · Catalog reviewed 2026-09-15
- Global · GuidanceOWASP Top 10 for Agentic Applications 2026OWASP GenAI Security Project · Catalog reviewed 2026-09-15
- Global · GuidanceOWASP Agent Control StandardOWASP GenAI Security Project · Catalog reviewed 2026-09-15
- EU · LawGeneral Data Protection RegulationEUR-Lex / European Union · Catalog reviewed 2026-09-15
- Global · GuidanceThe IIA AI Auditing FrameworkThe Institute of Internal Auditors · Catalog reviewed 2026-09-15
- EU · LawEU AI Omnibus — Regulation 2026/1744EUR-Lex / European Union · Catalog reviewed 2026-09-15
- EU · GuidanceEU AI Act implementation timelineEuropean Commission AI Act Service Desk · Catalog reviewed 2026-09-15
- EU · GuidanceAI Act transparency and enforcement announcementEuropean Commission · Catalog reviewed 2026-09-15
- US federal · PolicyOMB memorandum registerUS Office of Management and Budget · Catalog reviewed 2026-09-15
- US federal · PolicyOMB M-25-21 — Federal use of AIUS Office of Management and Budget · Catalog reviewed 2026-09-15
- US federal · PolicyOMB M-26-04 — Federal LLM procurement principlesUS Office of Management and Budget · Catalog reviewed 2026-09-15
- US Colorado · LawColorado SB26-189 — Automated decision-making technologyColorado General Assembly · Catalog reviewed 2026-09-15
- Global · GuidanceNIST AI Risk Management FrameworkUS National Institute of Standards and Technology · Catalog reviewed 2026-09-15
- Global · GuidanceNIST AI 600-1 — Generative AI ProfileUS National Institute of Standards and Technology · Catalog reviewed 2026-09-15
- Global · GuidanceNIST AI framework crosswalksNIST AI Resource Center · Catalog reviewed 2026-09-15
- Global · InitiativeNIST AI Agent Standards InitiativeUS National Institute of Standards and Technology · Catalog reviewed 2026-09-15
- Global · StandardISO/IEC 42001:2023 — AI management systemsInternational Organization for Standardization · Catalog reviewed 2026-09-15
- UK · PolicyUK Algorithmic Transparency Recording Standard scopeUK Government · Catalog reviewed 2026-09-15
- Canada · PolicyCanada automated decision-making directive amendmentsGovernment of Canada / Treasury Board · Catalog reviewed 2026-09-15
- Singapore · GuidanceSingapore Model AI Governance Framework for Agentic AIInfocomm Media Development Authority · Catalog reviewed 2026-09-15