DASHBOARD
Citizen Oversight Dashboard¶
Wireframes and layout specifications for the citizen-facing oversight interface.
Design Principle: Intentionally boring, legible, and civic. No dopamine UI. No panic UX.
1. HOME — Integrity at a Glance¶
Purpose¶
Answer in 10 seconds:
"Is this AI system behaving responsibly right now?"
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ MOBIUS ▸ Citizen Oversight [About] [FAQ] │
├─────────────────────────────────────────────────────────────┤
│ Mobius Integrity Index (MII) │
│ │
│ 🟢 0.97 — STABLE │
│ Integrity has remained stable for 42 days │
│ │
│ Thresholds: 🟢 Stable ≥0.95 🟡 Watch 0.90–0.94 🔴 <0.90│
├─────────────────────────────────────────────────────────────┤
│ [ Integrity Trend ] [ Intent Transparency ] [ Alerts ] │
│ │
│ 30d ▸ 90d ▸ 1y Declared: 1284 🔔 0 Critical │
│ ──────────── Executed: 1271 ⚠️ 2 Review │
│ ──────── Blocked: 13 ℹ️ 5 Info │
│ │
└─────────────────────────────────────────────────────────────┘
Design Intent¶
- One number
- One sentence
- No interpretation required
Header Strip (Always Visible)¶
| Element | Display |
|---|---|
| MII Value | Large, color-coded |
| Status | STABLE / WATCH / BREACH |
| Days Stable | Running count |
Status Colors¶
| Status | Color | Condition |
|---|---|---|
| Stable | 🟢 Green | MII ≥ 0.95 |
| Watch | 🟡 Yellow | MII 0.90–0.94 |
| Breach | 🔴 Red | MII < 0.90 |
2. ACTIVITY — What Changed?¶
Purpose¶
Answer:
"What actually happened, and why should I care?"
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ ACTIVITY LOG │
├─────────────────────────────────────────────────────────────┤
│ 🕘 Jan 26 — Policy Update Executed │
│ • Declared intent: Improve safety filter reliability │
│ • Risk level: Medium │
│ • Approvals: 1 human steward │
│ • Integrity impact: Neutral │
│ • Status: Completed │
│ │
│ [ Expand details ▾ ] │
├─────────────────────────────────────────────────────────────┤
│ 🕘 Jan 24 — Action Blocked │
│ • Reason: Missing justification │
│ • Integrity impact: Positive │
│ • Status: Prevented │
│ │
│ [ Expand details ▾ ] │
└─────────────────────────────────────────────────────────────┘
Key Rule¶
No code. No internal identifiers. Only: what / why / risk / outcome.
Event Fields¶
| Field | Description |
|---|---|
| Timestamp | When it happened |
| Summary | One-line description |
| Intent | Declared justification |
| Risk | Low / Medium / High |
| Impact | Integrity effect |
| Status | Completed / Blocked / Pending |
Filters¶
- By risk level
- By system component
- By date range
- By "blocked actions only"
3. OVERSIGHT — Early Warning, Not Panic¶
Purpose¶
Give citizens early warning, not panic.
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ OVERSIGHT SIGNALS │
├─────────────────────────────────────────────────────────────┤
│ Intent Coverage 🟢 Normal │
│ Review Latency 🟡 Slightly Elevated │
│ Repeated Overrides 🟢 None Detected │
│ Integrity Drift Risk 🟢 Low │
├─────────────────────────────────────────────────────────────┤
│ PATTERN WATCH │
│ • No recurring emergency actions │
│ • No decline in explanation quality │
│ • No acceleration of review bypass │
│ │
│ Confidence: High │
└─────────────────────────────────────────────────────────────┘
Signals Panel¶
| Signal | Description |
|---|---|
| Intent Coverage | Are actions being justified? |
| Review Latency | Are reviews happening on time? |
| Repeated Overrides | Are emergencies recurring? |
| Drift Risk | Is there a concerning trend? |
Pattern Watch¶
Shows patterns, not incidents:
- Repeated emergency overrides
- Declining explanation quality
- Increasing blocked actions
- Shortened review cycles
Each pattern has: - Explanation - Historical comparison - Confidence level
Prevention Goals¶
- Single-incident hysteria
- Social media pile-ons
- Reactionary conclusions
4. CITIZEN ACTION — Act, Carefully¶
Purpose¶
Oversight without mobs or vigilantism.
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ CITIZEN ACTION │
├─────────────────────────────────────────────────────────────┤
│ What would you like to do? │
│ │
│ [ View reports ] [ Subscribe ] [ Flag concern ] │
│ │
├─────────────────────────────────────────────────────────────┤
│ FLAG A CONCERN (Auditor) │
│ │
│ What are you seeing? [__________________________] │
│ Pattern or incident? ( ) Pattern ( ) Incident │
│ Why does this matter? [__________________________] │
│ │
│ ⚠️ Reports are recorded. Bad-faith use reduces trust score. │
│ │
│ [ Submit ] │
└─────────────────────────────────────────────────────────────┘
Action Options (Role-Based)¶
| Role | Actions |
|---|---|
| Observer | Read reports, subscribe, share |
| Auditor | Flag anomaly, attach evidence |
| Steward | Review escalations, vote on thresholds |
Important¶
Friction is intentional. Oversight ≠ outrage.
5. MEMORY — No Amnesia¶
Purpose¶
Prevent institutional amnesia.
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ SYSTEM MEMORY │
├─────────────────────────────────────────────────────────────┤
│ Total actions recorded: 18,402 │
│ Total blocked actions: 214 │
│ Integrity breaches: 2 (resolved) │
│ Longest stable period: 123 days │
├─────────────────────────────────────────────────────────────┤
│ HISTORICAL NOTES │
│ • "Review process strengthened after Jan 2025 incident" │
│ • "Thresholds adjusted after citizen audit" │
└─────────────────────────────────────────────────────────────┘
Memory Snapshot¶
| Metric | Description |
|---|---|
| Actions Recorded | Total logged actions |
| Blocked Actions | Prevention count |
| Breaches | With resolution status |
| Stable Period | Longest clean run |
Public Annotations¶
Citizens and stewards can attach notes:
"This change later caused X issue." "This was resolved after Y intervention."
Think public margin notes, not comments.
6. GOVERNANCE — Who Is Responsible?¶
Purpose¶
Accountability without harassment.
Wireframe¶
┌─────────────────────────────────────────────────────────────┐
│ GOVERNANCE MAP │
├─────────────────────────────────────────────────────────────┤
│ Decisions requiring human approval: │
│ • Production deployment │
│ • Integrity threshold changes │
│ • MIC mint/burn logic │
│ │
│ Oversight path: │
│ Citizen → Auditor → Steward → Public record │
│ │
│ No single entity has unilateral authority. │
└─────────────────────────────────────────────────────────────┘
Governance Map Contents¶
- What decisions require human approval
- What thresholds trigger escalation
- Who reviews what (roles, not names)
- How disputes are resolved
Clear Statement¶
"No single person or model has unilateral authority."
7. SAFETY FOOTER (Always Visible)¶
┌─────────────────────────────────────────────────────────────┐
│ Mobius does not monitor individuals. │
│ Mobius does not access private data. │
│ Mobius does not censor content. │
│ │
│ Mobius records behavior and justification — nothing more. │
└─────────────────────────────────────────────────────────────┘
This protects trust.
8. MOBILE MODE¶
Mobile view collapses into:
- MII status
- "What changed today?"
- Alerts (if any)
- One-tap "Learn more"
No dense charts on mobile.
First-Time User Prompt¶
Shown once, on first visit:
┌─────────────────────────────────────────────────────────────┐
│ │
│ Mobius does not tell you what to think. │
│ It shows you what is happening. │
│ │
│ Oversight works when citizens act carefully, not loudly. │
│ │
│ [ I understand ] │
│ │
└─────────────────────────────────────────────────────────────┘
Users must click to continue.
"A dashboard that lets citizens see institutional drift before it becomes irreversible."