Editorial Policy
Source hierarchy
Primary-source evidence takes precedence over second-hand reporting. In order of priority: official association announcements, primary regulatory/statutory text, manufacturers' own official newsroom pages, and certification registry change records outrank any secondary summary, press digest, or third-party recap of the same underlying fact.
AI is assistive only, human review is mandatory
AI may assist in drafting or surfacing candidate content, but AI output alone is never sufficient to publish. Every item published under the signals pipeline must pass a mandatory human review gate before it can reach a publicly visible "confirmed" state — this mirrors the database-level enforcement in the Signals capability's review-status state machine.
AI-generated content is disclosed to readers
Where AI played a leading role in producing a piece of content, that is disclosed to the reader explicitly — for example, labeled "AI-assisted, human-reviewed" — rather than presented as if it were produced entirely by a human editor.
News and compliance items require at least two reviewers
News and compliance-update entries require sign-off from at least two distinct reviewers before they are marked confirmed and shown publicly — this extends the Signals capability's reviewStatus state-transition table, which enforces this at the data layer, not just as an editorial convention.
Corrections are public and traceable
When published content is corrected, the correction history is never deleted — prior versions and the fact that a correction occurred remain visible and traceable, not silently overwritten. Adopted data corrections are listed in the corrections log.
Data corrections and data partnership or data-access requests: [email protected].
Advertising and PR content is clearly labeled
Sponsored, advertising, or PR-originated content is explicitly labeled and visually distinguished from independently sourced editorial content — readers are never left to guess which is which.
What payment can and cannot buy, and how to report a breach, is set out in our Neutrality policy.
How we weigh evidence
Every value we record carries an internal evidence grade: strong when the source states the value directly, moderate when it has to be read or derived from the source with some interpretation, and weak when the evidence is indirect or second-hand. When values conflict, the better-evidenced one is shown first and the other stays visible.
We no longer print the grade next to each value. The link to the source is the evidence: open it to see the document and the excerpt the value came from. How sources rank is described in the Methodology.