Identity and office integritySource coverage: partialCommunity input: unavailable
Trust and evidence
A score should never ask for blind trust.
Every factor, source, input type, confidence signal, response, and correction should remain visible.
Interactive concept. The application is in development.
Illustrative profile
Sample Representative
Member of Parliament · Sample Riding
Canada · Current sample termDigital ID plannedMethodology pending1 2 3 4
Illustrative factor breakdown
Voting and policy recordUnder human reviewCommunity input: unavailable
Promises and stated positionsInsufficient source coverageCommunity input: unavailable
Evidence quality and confidenceWeight to be publishedCommunity input: unavailable
AI-assisted record analysisAuthenticated civic sentimentUnverified inputAnonymous input
Correction history visible
Illustrative profile. All records, dates, and statuses are fictional.
01
The public-record layer.
AI-assisted analysis organizes source-backed conduct connected to the public office and term.
TRUST-01Trust & evidence
Formula inputs and decomposed outputs
Concept · planned- 1Source-backed public-record evaluation
- 1Published formula
- 2Versioned methodology
- 1Community input from RMR users
- 1Record analysis
- 2Authenticated civic sentiment
- 3Unverified input
- 4Anonymous input
- 5Evidence confidence
- 6Source coverage
02
The community-input layer.
User support, concern, ratings, comments, and contributed evidence remain separately identifiable.
TRUST-02Trust & evidence
Evidence packet anatomy
Concept · planned- Submitter claim
- Source links
- Relevant excerpts
- Dates
- Media references
- Office and term
- Context and limitations
- Human-review state
- Representative response
- Correction history
- Hash or signature reference
03
The Light Mathematics formula.
The methodology, factors, weights, exclusions, version, and confidence rules must be published before a score goes live.
Illustrative profile
Sample Representative
Member of Parliament · Sample Riding
Canada · Current sample termDigital ID plannedMethodology pending1 2 3 4
Illustrative factor breakdown
Identity and office integritySource coverage: partialCommunity input: unavailable
Voting and policy recordUnder human reviewCommunity input: unavailable
Promises and stated positionsInsufficient source coverageCommunity input: unavailable
Evidence quality and confidenceWeight to be publishedCommunity input: unavailable
AI-assisted record analysisAuthenticated civic sentimentUnverified inputAnonymous input
Correction history visible
Illustrative profile. All records, dates, and statuses are fictional.
04
Evidence travels with context.
Dates, sources, office, term, limitations, disputes, and review status remain attached.
TRUST-03Trust & evidence
AI record work versus human judgment
Concept · planned- Find sources
- Classify records
- Compare statements
- Flag contradictions
- Assess coverage
- Draft analysis
- Confirm context
- Submit civic signal
- Review high-impact claims
- Approve publication
- Resolve disputes
- Correct records
05
AI evaluates records. People make civic judgments.
The system may identify patterns or conflicts. It may not invent public opinion.
TRUST-05Trust & evidence
Evidence lifecycle
Concept · planned- 1Submitted
- 2Validated
- 3Human reviewed
- 4Published or disputed
- 5Responded or corrected
- 6Provenance anchored
- 7Correction branch returns to human review
06
Corrections extend history.
Responses and corrections create new visible events rather than erasing the original record.
PROFILE-02Profiles
Public-role identity linkage
Concept · plannedFictional public-role identity
- Person
- Office
- Jurisdiction
- District
- Term
- Official identifiers
- Public actions
- Evidence
- Responses
- Corrections
Keep the record moving
Inspect the methodology