AI Use and Editorial Standards
Last updated: 31 August 2026
This statement describes how Briefing Items in the Stratsec Radar are produced and the standards applied to them. It forms part of the Terms of Use. Defined terms have the meanings given in section 1 of those Terms.
1. What AI does
AI models supplied by different providers monitor regulators, standards bodies, legislatures and official journals across the four jurisdictions the Radar covers. Each development they identify is screened for relevance against the sixteen coverage domains. Developments that pass that screen are drafted into a first version of a Briefing Item in the five-field format.
Drafts are then cross-reviewed. A draft produced by one provider’s model is critiqued against its primary source by a model from a different provider, and where that critique identifies a defect the draft is revised before any person reads it.
2. What AI does not do
AI does not publish, does not determine which developments matter, does not form a view on your exposure, and does not exercise the editorial judgment for which a named person is accountable.
Every Briefing Item reaches a human editor as a draft and remains a draft until a person approves it.
3. Human review
A named analyst reads each Briefing Item against its primary source and approves it before publication, and former Chief Information Security Officers review material in the sectors in which they held that role. This applies to every Briefing Item, and not to a sample or only to items a model has marked as uncertain.
The reviewers address what a requirement will cost an organisation to implement, and not only what the text of the instrument says. Reviewers are former CISOs of Forbes Global 2000 companies and former cybersecurity practice leaders from the Big Four, IBM and Accenture.
Analysts and reviewers are available to subscribers directly, and a question arising from a Briefing Item may be put to the person who approved it.
4. Model providers, and what is sent to them
Models supplied by Anthropic, OpenAI and Google are used in the production of the Radar. Assignments between providers are revised as model capability changes, and no part of the method depends on a single supplier.
The material sent to those providers is published source content: the text of an instrument, a consultation document, a supervisory statement or a standards publication. Where such a source names an individual, for example a named official or the subject of an enforcement decision, that name forms part of the published source and is processed as published material.
No subscriber personal data is sent to any model provider. Account details, Member Lens configuration, the triage state recorded by your team in the Portal, and the content of any call with an analyst or a reviewer are held within the service and are never used as model input.
These providers do not process personal data on our behalf, and they are therefore not engaged as processors of subscriber data. The Privacy Policy lists them as technology suppliers, separately from the providers that do handle personal data, which are the hosting, email, consent and payment providers.
5. Sources
Primary sources take precedence: the instrument itself, the delegated act, the consultation document, the regulator’s own announcement, or the standards body’s published text. Every Briefing Item links to its primary source. Where that source requires a subscription or payment for access, the Briefing Item states this.
Journalism is used for context and is never the sole source for a statement about a legal obligation. Other newsletters are not used as sources.
6. Status and scope
Each Briefing Item states whether an instrument is proposed, adopted, in force or applicable, gives the relevant date, and identifies which of those the date refers to.
Where an instrument is a directive requiring national transposition, the Briefing Item identifies it as such, because dates and obligations vary between member states and a statement about the European Union as a whole can be inaccurate in any one of them.
Each Briefing Item identifies who falls within scope, using the categories the instrument itself uses.
7. Uncertainty
Each Briefing Item includes a field recording what remains unclear. That field is not left empty for presentational reasons.
Where a question is unsettled, the Briefing Item states what is unclear and what would resolve it, so that a subscriber can judge whether to act or to wait for guidance.
8. What is not done
- No Briefing Item is published without human review and approval.
- No source, citation or quotation is generated. Material that cannot be linked to a primary source is not published.
- No representation is made that coverage of any domain or jurisdiction is complete.
- No product and no vendor is recommended.
- No legal advice is given.
9. Corrections
Where an error is identified in a published Briefing Item, the correction is made in that item, dated and described. Items are not amended silently.
Errors may be reported to [email protected], and a report from a subscriber is treated as a priority.
10. Contact
Post-Quantum Institute, doing business as Stratsec
155 N Wacker Dr, Suite 4250
Chicago, IL 60606
United States
[email protected]