Wiki source code of Governance

Version 2.1 by Robert Schaub on 2025/12/14 18:59

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