Foundations

Observed Index

The Observed Index is ETHORITY’s reality layer: a global, append-only record of observable facts about AI systems and related artifacts. It does not score, endorse, or interpret.

What it is

A durable record of statements like: “this system exists”, “this artifact was published”, “this event occurred at time T”. The Observed Index is deliberately boring — it becomes reliable precisely because it refuses to judge.

Observed sources & collectors

All automated collectors operate under the Observed Sources Contract. Collectors record what was observed, where, and when — but never interpret, score, certify, or assign trust.

For engineers

Integrate a neutral source of truth without embedding regulatory or policy logic into your codebase.

Scenario: an agent checks an observed registry record before calling an external API.

For regulators & auditors

Cite observable statements without endorsing the system or applying policy prematurely.

Scenario: a published claim is referenced as evidence in a dispute or investigation.

What it is not

  • • Not a leaderboard
  • • Not a compliance certificate
  • • Not a safety claim
  • • Not an “ethical AI” stamp

Next step

Once facts are observed, governance systems may interpret them under explicit rule sets — without modifying the observed layer itself.

Continue: Governance Layer →