EkoH

EkoH (Expertise + Ethics Ledger) — the merit ledger used by Kollective Intelligence and consumed by Smart Vote and other ranking/visibility surfaces.

EkoH does not produce “decision readings.” Readings (lenses) are a Smart Vote concern.
EkoH produces auditable weight inputs: domain expertise vectors, ethics multipliers, privacy constraints, and immutable snapshots.

---

1) What EkoH is responsible for

Core outputs

• Domain expertise vectors per user, aligned to a formal taxonomy (recommended: UNESCO ISCED-F).
• Ethics multiplier per user (governance score applied by Smart Vote per lens policy).
• Confidentiality level (public / pseudonym / anonymous) controlling disclosure.
• Audit artifacts: score history, context adjustment logs, integrity events.
• Immutable snapshots bound to consultations (Smart Vote consumes snapshots, not “live” scores).

What EkoH is not

• Not the voting engine.
• Not the reading/lens engine.
• Not the consultation orchestration layer. Those live under Smart Vote / ethiKos.

---

2) Ledger primitives (conceptual model)

EkoH stores merit as two orthogonal components:

1. Expertise score per domain
   A user can be strong in one domain and neutral in another; influence is domain-matched.

2. Ethics multiplier (global)
   A governance score that modulates influence according to integrity signals and platform rules.
   How ethics is applied to voting weights is lens-defined in Smart Vote.

---

3) Canonical storage model (schema: ekoh_smartvote)

Core tables used to store expertise, ethics, privacy, and audit context:

Table / Model	Purpose	Key fields (illustrative)
expertise_category	Domain taxonomy (e.g., UNESCO ISCED-F).	code, name, parent_id, depth, path
user_expertise_score	Per-user, per-domain expertise score.	user_id, category_id, raw_score, weighted_score
user_ethics_score	Per-user ethics multiplier.	user_id, ethical_score
confidentiality_setting	Privacy level for disclosure.	user_id, level (public/pseudonym/anonymous)
score_configuration	Versioned coefficients/knobs.	weight_name, weight_value, field (nullable)
context_analysis_log	Explainable adjustments / context traces.	entity_type, entity_id, input_metadata (json), adjustments_applied (json)
score_history	Immutable audit trail for score changes.	merit_score_id, old_value, new_value, change_reason, created_at
integrity_event	Integrity anomalies & handling state.	event_type, score (json), handled, created_at

Optional (only if you need stakeholder/cohort analytics):

• demographic_attribute, demographic_choice, user_demographic

---

4) Scoring rules (high-level)

EkoH expertise scoring commonly uses three axes (coefficients are configuration-driven):

• Quality (accuracy, evidence, peer signals)
• Expertise (verified credentials / demonstrated competence)
• Frequency / Recency (continued engagement, decay)

Illustrative aggregation:

• weighted_score = Q*RAW_WEIGHT_QUALITY + X*RAW_WEIGHT_EXPERTISE + F*RAW_WEIGHT_FREQUENCY

All coefficient values must live in score_configuration and be versioned.

---

5) Ethics multiplier (governance score)

• Ethics is treated as a governance signal tied to platform rules and evidence.
• Any negative adjustment must be:
  • logged (reason codes + evidence pointers),
  • auditable,
  • appealable (process is policy-defined).

Typical policy bounds are configuration-driven (examples only):

• floor (e.g., 0.20)
• cap (e.g., 1.50)

---

6) Snapshotting (critical for Smart Vote)

Smart Vote consumes immutable EkoH snapshots bound to a consultation:

• snapshot references the exact expertise vectors + ethics multipliers used
• snapshot references the score_configuration version hash
• snapshot is the reproducibility anchor for audits and result replays

---

7) Smart Vote integration (what Smart Vote reads from EkoH)

Smart Vote takes:

• S(u,d) = user expertise score in domain d (from user_expertise_score)
• E(u) = ethics multiplier (from user_ethics_score)
• confidentiality constraints (from confidentiality_setting)
• snapshot manifest (bound to consultation)

Then applies a lens-defined weight formula, for example:

• W(u,c) = E(u) × ( 1 + Σ_d ( S(u,d) × R(c,d) ) )
  or an alternative lens preserving an unscaled baseline vote.

EkoH’s role is to make those inputs legible, auditable, and contestable.

---

Summary

EkoH is a merit ledger: domain expertise vectors + ethics multipliers + privacy + audit context, with immutable snapshots for governance-grade reproducibility.
Smart Vote is the decision engine that applies those inputs to modalities and publishes readings.