Wiki source code of Governance

Last modified by Robert Schaub on 2025/12/24 21:52

Show last authors
1 = Governance =
2 FactHarbor governance is **simple, transparent, flat, and focused on enabling automation**.
3 {{include reference="FactHarbor.Organisation.Diagrams.Governance Structure.WebHome"/}}
4 == 1. Governance Philosophy ==
5 **Core Principles**:
6 * **Automation over bureaucracy**: Minimize manual processes
7 * **Transparency by default**: Open decision-making
8 * **Community input**: Listen but decide decisively
9 * **Measured outcomes**: Data drives decisions
10 * **Adaptive structure**: Evolve as needed
11 == 2. Organizational Structure ==
12 **Flat Cooperative Model**:
13 * Small organization with collaborative teamwork
14 * No hierarchical management layers
15 * Decisions by consensus when possible, voting when needed
16 * Everyone contributes across multiple areas
17 **General Assembly** → **Governing Team** → **Team Members**
18 === 2.1 General Assembly (All Members) ===
19 Decides: Governing Team election, statutes, major strategic changes, budget
20 Meets: Annually
21 === 2.2 Governing Team ===
22 Decides: Strategy, policy, budget allocation, hiring
23 Meets: Quarterly
24 Size: small group (Facilitator, Coordinator, Treasurer + others)
25 === 2.3 Team Members ===
26 * **Technical Coordinator**: AKEL & infrastructure
27 * **Community Coordinator**: Moderators & contributors
28 * **Moderators** (part-time): Handle abuse/disputes
29 == 3. Decision Authority ==
30 **Day-to-day**: Technical Coordinator + Community Coordinator
31 **Tactical**: Governing Team (simple majority)
32 **Strategic**: General Assembly (2/3 majority)
33 == 4. Policy Development ==
34 **RFC Process** (Request for Comments):
35 1. Anyone drafts proposal
36 2. Community discussion
37 3. Governing Team reviews and votes
38 4. Decision published
39 **Emergency changes**: Technical Coordinator can act immediately, Governing Team ratifies later
40 == 5. Financial Governance ==
41 * Annual budget approved by General Assembly
42 * Two-signature requirement for >CHF 5,000
43 * Governing Team approval for >CHF 20,000
44 * Quarterly financial reports
45 * Annual independent audit
46 == 6. Automation Governance ==
47 **Core Principle**: AKEL makes content decisions. Humans make system decisions.
48 === 6.1 Decision Boundary ===
49 **What AKEL Decides (Automated)**:
50 * All claim verdicts and confidence scores
51 * All evidence assessments and relevance scores
52 * All source track record scores
53 * All risk tier classifications
54 * All publication decisions
55 * All scenario extractions
56 **Rationale**: These decisions must be automated for scale, consistency, transparency, and to avoid human bias. Humans cannot process millions of claims reliably.
57 **Human Role**: Monitor aggregate performance metrics, identify systematic issues, improve algorithms.
58 **What Humans Decide**:
59 **Strategic Decisions** (General Assembly, 2/3 majority):
60 * Mission and values
61 * Risk tier policy definitions
62 * Major architectural changes
63 * Budget allocation
64 * Dissolution
65 **Tactical Decisions** (Governing Team, simple majority):
66 * Algorithm parameter ranges (within policy)
67 * Infrastructure investments
68 * Hiring and role assignments
69 * Community policies
70 * Partnership agreements
71 **Operational Decisions** (Domain Owners, autonomous):
72 * Technical Coordinator: AKEL performance optimizations, infrastructure changes
73 * Community Coordinator: Community process improvements, documentation
74 * Moderators: Handling AKEL-flagged items, detection improvement proposals
75 **Emergency Decisions** (any team member, ratified by Governing Team):
76 * Critical security issues
77 * Legal compliance requirements
78 * Immediate safety concerns
79 === 6.2 Principle: Fix the System, Not the Data ===
80 **When AKEL makes a "wrong" decision**:
81 * ❌ Do NOT manually override that specific verdict
82 * ✅ DO investigate: Is this a systematic issue?
83 * ✅ DO improve: Change algorithm/policy to handle such cases better
84 * ✅ DO test: Validate improvement on historical data
85 * ✅ DO deploy: Roll out improved system
86 * ✅ DO monitor: Check if metrics improve
87 **Example**:
88 * Bad: "AKEL rated this source too low, I'll manually boost it"
89 * Good: "AKEL consistently under-rates peer-reviewed sources. Let's adjust the scoring algorithm to weight peer-review more heavily."
90 === 6.3 Governance of AKEL ===
91 **Quarterly Performance Review**:
92 * Who: Governing Team + Technical Coordinator
93 * What: Review AKEL performance metrics, bias audits, user feedback patterns
94 * Output: Performance report, improvement priorities, policy updates if needed
95 **Performance Metrics Monitored**:
96 * Processing speed (P50, P95, P99)
97 * Success rate and error rate
98 * Evidence completeness
99 * Confidence score distribution
100 * User feedback (helpful/unhelpful ratio)
101 * Bias indicators (by domain, source type, etc.)
102 **Triggers for Policy Review**:
103 * Metrics outside acceptable ranges
104 * Systematic bias detected
105 * Major user complaints about fairness
106 * Legal/compliance concerns
107 * New domains requiring special handling
108 **Algorithm Change Process**:
109 1. Identify issue from metrics
110 2. Propose solution (RFC - Request for Comments)
111 3. Test in staging environment
112 4. Measure impact on historical data
113 5. Technical Coordinator approves (or escalates to Governing Team for policy changes)
114 6. Deploy with monitoring
115 7. Evaluate results
116 === 6.4 Human Intervention Criteria ===
117 **Legitimate reasons to intervene**:
118 * ✅ AKEL explicitly flags item for human review
119 * ✅ System metrics show performance degradation
120 * ✅ Legal/safety issue requires immediate action
121 * ✅ User reports reveal systematic bias pattern
122 **Illegitimate reasons** (system improvement needed instead):
123 * ❌ "I disagree with this verdict" → Improve algorithm
124 * ❌ "This source should rank higher" → Improve scoring rules
125 * ❌ "Manual quality gate before publication" → Defeats purpose of automation
126 * ❌ "I know better than the algorithm" → Then improve the algorithm
127 === 6.5 Consent-Based Decision Making ===
128 **For system changes**, use consent not consensus (from Sociocracy 3.0):
129 **Consent** = No principled objections
130 * Faster than consensus
131 * Respects concerns without requiring full agreement
132 * "I can live with this and support it"
133 **Process**:
134 1. Proposal presented (RFC)
135 2. Clarifying questions
136 3. Reactions and concerns
137 4. Proposer integrates feedback
138 5. Consent round: Any principled objections?
139 6. If no objections → Decision made
140 7. If objections → Integrate and repeat
141 **Use for**:
142 * Algorithm changes
143 * Policy updates
144 * Infrastructure investments
145 * Process changes
146 **Not for**:
147 * Strategic decisions (use voting)
148 * Emergency decisions (use autonomous authority)
149 == 7. Transparency & Accountability ==
150 **Always public**: Policies, structure, board membership, financials, quality metrics, major decisions
151 **Published quarterly**: Activity reports, metrics, moderation stats
152 **Internal documentation**: All meetings, decisions, actions retained indefinitely
153 == 7. Conflict Resolution ==
154 **User disputes**: Moderator → Appeal to different moderator → Governing Team (final)
155 **Timeline**: Reasonable timeframe
156 == 8. Moderation Oversight ==
157 **Moderator requirements**: high reputation, 6+ months active, clean record
158 **Review**: Quarterly by Community Coordinator, annually by Governing Team
159 **Appeal**: Any user can appeal promptly
160 == 9. Code of Conduct ==
161 **Governing Team**: Act in org interest, disclose conflicts, maintain confidentiality
162 **Team Members**: Follow procedures, professional conduct, document decisions
163 **Moderators**: Impartial decisions, respect privacy, respond timely
164 == 10. Amendment Process ==
165 **Minor changes**: Governing Team decision
166 **Major changes**: General Assembly (2/3 vote)
167 **Emergency**: Governing Team can act, must ratify at next Assembly
168 == 11. Related Pages ==
169 * [[Organisational Model>>FactHarbor.Organisation.Organisational-Model]]
170 * [[Legal Framework>>FactHarbor.Organisation.Legal-Framework]]
171 * [[Transparency Policy>>FactHarbor.Organisation.How-We-Work-Together.Transparency-Policy]]