Skip to Content

Wiki source code of Governance

Version 3.2 by Robert Schaub on 2025/12/16 20:25
Show last authors
1 = Governance =
2
3 FactHarbor is governed collaboratively with clear separation between **organizational policy and decisions** and **technical implementation**.
4
5 == Governance Structure ==
6
7 {{include reference="Test.FactHarborV09.Organisation.Diagrams.Governance Structure.WebHome"/}}
8
9 * **Governing Team** – Sets high-level policy, organizational direction, funding priorities
10 * **Lead** – Coordinates execution, represents organization publicly
11 * **Core Maintainers** – Technical and specification decisions, code/spec review
12 * **Domain Experts** – Subject-matter authority in specialized areas
13 * **Community Contributors** – Feedback, proposals, and participation in decision-making
14
15 ----
16
17 == Decision-Making Levels ==
18
19 === Technical Decisions (Maintainers) ===
20
21 **Scope**: Architecture, data model, AKEL configuration, quality gates, system performance
22
23 **Process**:
24 * Proposals discussed in technical forums
25 * Review by core maintainers
26 * Consensus-based approval
27 * Breaking changes require broader community input
28 * Quality gate adjustments require rationale and audit validation
29
30 **Examples**:
31 * Adding new quality gate
32 * Adjusting AKEL parameters
33 * Modifying audit sampling algorithms
34 * Database schema changes
35
36 === Policy Decisions (Governing Team + Community) ===
37
38 **Scope**: Risk tier policies, publication rules, content guidelines, ethical boundaries
39
40 **Process**:
41 * Proposal published for community feedback
42 * Discussion period (recommendation: minimum 14 days for major changes)
43 * Governing Team decision with community input
44 * Transparency in reasoning
45 * Risk tier policy changes require Expert consultation
46
47 **Examples**:
48 * Defining Tier A domains
49 * Setting audit sampling rates
50 * Content moderation policies
51 * Community guidelines
52
53 === Domain-Specific Decisions (Experts) ===
54
55 **Scope**: Domain quality standards, source reliability in specialized fields, Tier A content validation
56
57 **Process**:
58 * Expert consensus in domain
59 * Documented reasoning
60 * Review by other experts
61 * Escalation to Governing Team if unresolved
62 * Experts set domain-specific audit criteria
63
64 **Examples**:
65 * Medical claim evaluation standards
66 * Legal citation requirements
67 * Scientific methodology thresholds
68 * Tier A approval criteria by domain
69
70 ----
71
72 == AI and Human Roles in Governance ==
73
74 === Human-Only Governance Decisions ===
75
76 The following can **never** be automated:
77
78 * **Ethical boundary setting** – What content is acceptable, what harm thresholds exist
79 * **Risk tier policy** – Which domains are Tier A/B/C (though AKEL can suggest)
80 * **Audit system oversight** – Quality standards, sampling strategies, auditor selection
81 * **Dispute resolution** – Conflicts between experts, controversial decisions
82 * **Community guidelines enforcement** – Bans, suspensions, conflict mediation
83 * **Organizational direction** – Mission, vision, funding priorities
84
85 === AKEL Advisory Role ===
86
87 AKEL can **assist but not decide**:
88
89 * Suggest risk tier assignments (humans validate)
90 * Flag content for expert review (humans decide)
91 * Identify patterns in audit failures (humans adjust policy)
92 * Propose quality gate refinements (maintainers approve)
93 * Detect emerging topics needing new policies (Governing Team decides)
94
95 === Transparency Requirement ===
96
97 All governance decisions must be:
98 * **Documented** with reasoning
99 * **Published** for community visibility
100 * **Reviewable** by community members
101 * **Reversible** if evidence of error or harm
102
103 ----
104
105 == Audit System Governance ==
106
107 === Audit Oversight Committee ===
108
109 **Composition**: Maintainers, Domain Experts, and Governing Team member(s)
110
111 **Responsibilities**:
112 * Set quality standards for audit evaluation
113 * Review audit statistics and trends
114 * Adjust sampling rates based on performance
115 * Approve changes to audit algorithms
116 * Oversee auditor selection and rotation
117 * Publish transparency reports
118
119 **Meeting Frequency**: Recommendation: Regular meetings as needed
120
121 **Reporting**: Recommendation: Periodic transparency reports to community
122
123 === Audit Performance Metrics ===
124
125 Tracked and published:
126 * Audit pass/fail rates by tier
127 * Common failure patterns
128 * System improvements implemented
129 * Time to resolution for audit failures
130 * Auditor performance (anonymized)
131
132 === Feedback Loop Governance ===
133
134 **Process**:
135 1. Audits identify patterns in AI errors
136 2. Audit Committee reviews patterns
137 3. Maintainers propose technical fixes
138 4. Changes tested in sandbox
139 5. Community informed of improvements
140 6. Deployed with monitoring
141
142 **Escalation**:
143 * Persistent high failure rates → Pause AI publication in affected tier/domain
144 * Critical errors → Immediate system review
145 * Pattern of harm → Policy revision
146
147 ----
148
149 == Risk Tier Policy Governance ==
150
151 === Risk Tier Assignment Authority ===
152
153 * **AKEL**: Suggests initial tier based on domain, keywords, content analysis
154 * **Moderators**: Can override AKEL for individual content
155 * **Experts**: Set tier policy for their domains
156 * **Governing Team**: Approve tier policy changes, resolve tier disputes
157
158 === Risk Tier Review Process ===
159
160 **Triggers for Review**:
161 * Significant audit failures in a tier
162 * New emerging topics or domains
163 * Community flags systematic misclassification
164 * Expert domain recommendations
165 * Periodic policy review
166
167 **Process**:
168 1. Expert domain review (identify if Tier A/B/C appropriate)
169 2. Community input period (recommendation: sufficient time for feedback)
170 3. Audit Committee assessment (error patterns in current tier)
171 4. Governing Team decision
172 5. Implementation with monitoring period
173 6. Transparency report on rationale
174
175 === Current Tier Assignments (Baseline) ===
176
177 **Tier A**: Medical, legal, elections, safety/security, major financial decisions
178
179 **Tier B**: Complex science causality, contested policy, historical interpretation with political implications, significant economic impact
180
181 **Tier C**: Established historical facts, simple definitions, well-documented scientific consensus, basic reference info
182
183 **Note**: These are guidelines; edge cases require expert judgment
184
185 ----
186
187 == Quality Gate Governance ==
188
189 === Quality Gate Modification Process ===
190
191 **Who Can Propose**: Maintainers, Experts, Audit Committee
192
193 **Requirements**:
194 * Rationale based on audit failures or system improvements
195 * Testing in sandbox environment
196 * Impact assessment (false positive/negative rates)
197 * Community notification before deployment
198
199 **Approval**:
200 * Technical changes: Maintainer consensus
201 * Policy changes (e.g., new gate criteria): Governing Team approval
202
203 **Examples of Governed Changes**:
204 * Adjusting contradiction search scope
205 * Modifying source reliability thresholds
206 * Adding new bubble detection patterns
207 * Changing uncertainty quantification formulas
208
209 ----
210
211 == Community Participation ==
212
213 === Open Discussion Forums ===
214
215 * Technical proposals (maintainer-led)
216 * Policy proposals (Governing Team-led)
217 * Domain-specific discussions (Expert-led)
218 * Audit findings and improvements (Audit Committee-led)
219
220 === Proposal Mechanism ===
221
222 Anyone can propose:
223 1. Submit proposal with rationale
224 2. Community discussion (recommendation: minimum timeframe for feedback)
225 3. Relevant authority reviews (Maintainers/Governing Team/Experts)
226 4. Decision with documented reasoning
227 5. Implementation (if approved)
228
229 === Transparency ===
230
231 * All decisions documented in public wiki
232 * Audit statistics published periodically
233 * Governing Team meeting minutes published
234 * Expert recommendations documented
235 * Community feedback acknowledged
236
237 ----
238
239 == Dispute Resolution ==
240
241 === Conflict Between Experts ===
242
243 1. Experts attempt consensus
244 2. If unresolved, escalate to Governing Team
245 3. Governing Team appoints neutral expert panel
246 4. Panel recommendation
247 5. Governing Team decision (final)
248
249 === Conflict Between Maintainers ===
250
251 1. Discussion in maintainer forum
252 2. Attempt consensus
253 3. If unresolved, Lead makes decision
254 4. Community informed of reasoning
255
256 === User Appeals ===
257
258 Users can appeal:
259 * Content rejection decisions
260 * Risk tier assignments
261 * Audit outcomes
262 * Moderation actions
263
264 **Process**:
265 1. Submit appeal with evidence
266 2. Reviewed by independent moderator/expert
267 3. Decision with reasoning
268 4. Final appeal to Governing Team (if warranted)
269
270 ----
271
272 == Related Pages ==
273
274 * [[AKEL (AI Knowledge Extraction Layer)>>FactHarbor.Specification.AI Knowledge Extraction Layer (AKEL).WebHome]]
275 * [[Automation>>FactHarbor.Specification.Automation.WebHome]]
276 * [[Requirements (Roles)>>FactHarbor.Specification.Requirements.WebHome]]
277 * [[Organisational Model>>FactHarbor.Organisation.Organisational-Model]]