How reliable is this source, in this scope?
Identity, historical accuracy, authority, provenance, timeliness, independence, and manipulation risk.
Explainable intelligence
Source trust is not event confidence.
GSIP evaluates who produced information and separately evaluates whether a specific event claim is supported. That distinction prevents one famous source, one social post, or ten syndicated copies from becoming false certainty.
Loading methodology version…
How well supported is this exact claim, right now?
Independent corroboration, contradictions, official confirmation, recency, geolocation, and supporting evidence.
Source taxonomy
Eight categories, evaluated independently from trust
Category describes how information was obtained. Trust tier describes how much weight a particular source receives.
Trust framework
Five controlled evidence tiers
Tiers are assigned by GSIP policy and reviewers. The recommendation engine may consume them, but cannot assign or override them.
Loading trust taxonomy…
Calculation input
Event confidence factors
Integrity rules
Non-negotiable safeguards
Live source projection
Tier A sources: 0Sources currently referenced by canonical events
SourceCategoryReliabilityScopeEvents
No Tier A sources appear in the current canonical event window.
Published-media directory
Every publisher GSIP read, named in full
The complete list of news publishers whose reporting entered the latest completed collection cycle, with the number of distinct provider records each one contributed. It records what was read, not what is trusted: a publisher listed here carries no trust tier of its own, and inclusion is never an endorsement.
Loading the published-media directory…
Publisher domainObservationsCountry
Loading the published-media directory…
Reading a publisher is not trusting it. Publisher trust and per-claim event confidence are assessed separately, and no single publisher can confirm an event on its own.