How Civic Ledger turns government activity into civic intelligence.

1. Observe

Collect attributable government and public records. Preserve the source, publication date and relevant context rather than treating an AI summary as the original record.

2. Resolve

Map the institution, jurisdiction, authority, office and event. Keep uncertain matches visible instead of silently merging different institutions or people.

3. Interpret carefully

Break ambiguous terminology into specific questions. Identify the actor, action, instrument, scope, timeframe and conditions before assigning meaning.

4. Evaluate narrowly

Ask bounded machine questions that can be checked against evidence. A classification or routing decision should not become an unsupported judgment about a person's political character.

5. Represent uncertainty

These states preserve the difference between contrary evidence, missing evidence and a question that remains unsettled. Uncertainty should not disappear merely because an interface needs an answer.

6. Keep the public evidence layer neutral

The same public evidence exists before application of a person's private Civic Profile. A sponsor's preference or participant's values must not rewrite an observed public act.

7. Apply private relevance

Compare public activity with participant-supplied priorities and rules only within that person's agent. Private preferences remain separate from the shared institutional record and sponsor-facing outputs.

8. Apply authority

Alert the participant, request confirmation or execute a narrow explicit standing instruction. Label delegated execution and preserve its instruction provenance. A fresh human confirmation is a different response state.

9. Aggregate only authorized mission information

A Civic Pulse processes only the information covered by that mission's Participant Authorization. Apply the disclosed minimization, de-identification and privacy thresholds before releasing the defined aggregate product.

10. Maintain provenance

Make source and methodology visible. Record the relevant versions, correction history and response provenance so findings can be assessed with their context and limits. A sample is not automatically representative of everyone in a jurisdiction.

11. Preserve Observed versus Prototype State

Keep actual public action separate from experimental alternatives. A prototype identifies its observed baseline, proposed change and demonstrated result. Only later official adoption can change the observed government record.

Read the technical architecture appendix · Explore Mirror State