Skip to Content

Wiki source code of Roles & Publication Modes

Last modified by Robert Schaub on 2026/02/08 08:27
Show last authors
1 = Roles & Publication Modes =
2
3 **Version:** 0.9.70
4 **Last Updated:** December 21, 2025
5 **Status:** CORRECTED - Automation Philosophy Consistent
6
7 This page defines user roles and publication modes in FactHarbor.
8
9
10 == 1. User Roles ==
11
12 === 1.1 Contributors ===
13
14 **Who**: Community members who suggest system improvements
15
16 **Responsibilities:**
17
18 * Participate in sampling audits (analyze patterns, not individual outputs)
19 * Suggest algorithm/prompt improvements based on findings
20 * Document systematic issues observed
21 * Contribute to system improvement discussions
22
23 **Can:**
24
25 * Edit published content (changes apply immediately, Wikipedia-style)
26 * Flag quality issues (for sampling audit)
27 * Earn reputation through contributions
28 * Participate in RFC (Request for Comments) processes
29
30 **Cannot:**
31
32 * Approve content before publication
33 * Override quality gates
34 * Act as gatekeepers
35 * Manually fix individual AI outputs
36
37 **Note**: Contributors improve THE SYSTEM, not individual outputs.
38
39
40 === 1.2 Trusted Contributors ===
41
42 **Who**: Contributors with proven track record and domain expertise
43
44 **Responsibilities:**
45
46 * Same as Contributors, plus:
47 * Review complex algorithm changes
48 * Provide domain expertise on contested claims
49 * Mentor new contributors
50
51 **Can:**
52
53 * Everything Contributors can do, plus:
54 * Participate in higher-level system design decisions
55 * Review RFC proposals
56 * Access more detailed system metrics
57
58 **Cannot:**
59
60 * Approve content before publication
61 * Override quality gates
62 * Act as gatekeepers
63
64 **Important**: "Trusted" refers to judgment quality, NOT approval authority.
65
66
67 === 1.3 Moderators ===
68
69 **Who**: Team members focused on community health and abuse prevention
70
71 **Responsibilities:**
72
73 * Handle abuse, spam, and harassment
74 * Enforce community guidelines
75 * Respond to user reports
76 * Manage bans and appeals
77
78 **Can:**
79
80 * Hide abusive content
81 * Ban users for policy violations
82 * Review appeals
83 * Escalate serious issues to Governing Team
84
85 **Cannot:**
86
87 * Approve content for publication
88 * Review content quality before publication
89 * Override quality gates for content
90 * Act as editorial gatekeepers
91
92 **Critical Distinction:**
93 ```
94 Moderators handle: ABUSE (spam, harassment, violations)
95 Moderators DO NOT handle: CONTENT QUALITY (that's automated)
96 ```
97
98
99 === 1.4 Domain Trusted Contributors (Optional, Task-Specific) ===
100
101 **Who**: Subject matter specialists invited for specific high-stakes disputes
102
103 **Not a permanent role**: Contacted externally when needed for contested claims in their domain
104
105 **When used:**
106
107 * Medical claims with life/safety implications
108 * Legal interpretations with significant impact
109 * Scientific claims with high controversy
110 * Technical claims requiring specialized knowledge
111
112 **Process:**
113
114 * Moderator identifies need for expert input
115 * Contact expert externally (don't require them to be users)
116 * Trusted Contributor provides written opinion with sources
117 * Opinion added to claim record
118 * Trusted Contributor acknowledged in claim
119
120 **Important**: This is CONSULTATION, not APPROVAL. Their opinion is added to the evidence, not used as a gate.
121
122 **User Needs served**: UN-16 (Expert validation status)
123
124
125 == 2. Publication Modes ==
126
127 **Fulfills**: UN-1 (Trust indicators), UN-16 (Review status transparency)
128
129 FactHarbor uses **TWO publication modes** (not three). Focus is on transparency and confidence scoring, not gatekeeping.
130
131 === 2.1 Mode 1: Draft-Only ===
132
133 **Status**: Not visible to public
134
135 **When Used:**
136
137 * Quality gates failed
138 * Confidence below threshold
139 * Structural integrity issues
140 * Insufficient evidence
141
142 **What Happens:**
143
144 * Content remains private
145 * System logs failure reasons
146 * Prompts/algorithms improved based on patterns
147 * Content may be re-processed after improvements
148
149 **NOT "pending human approval"** - it's blocked because it doesn't meet automated quality standards.
150
151
152 === 2.2 Mode 2: AI-Generated (Public) ===
153
154 **Status**: Published and visible to all users
155
156 **When Used:**
157
158 * Quality gates passed
159 * Confidence ≥ threshold
160 * Meets structural requirements
161 * Sufficient evidence found
162
163 **Includes:**
164
165 * Confidence score displayed (0-100%)
166 * Risk tier badge (A/B/C)
167 * Quality indicators
168 * Clear "AI-Generated" labeling
169 * Sampling audit status
170
171 **Labels by Risk Tier:**
172
173 * **Tier A (High Risk):** "⚠️ AI-Generated - High Impact Topic - Seek Professional Advice"
174 * **Tier B (Medium Risk):** "🤖 AI-Generated - May Contain Errors"
175 * **Tier C (Low Risk):** "🤖 AI-Generated"
176
177 **User Contributions:**
178
179 * User edits apply immediately (Wikipedia model)
180 * All changes logged and versioned
181 * May be selected for sampling audit
182 * Reputation earned for quality contributions
183
184 === REMOVED: Mode 3 ===
185
186 **V0.9.50 Decision:** No centralized approval workflow.
187
188 **Rationale:**
189
190 * Defeats automation purpose
191 * Creates bottleneck
192 * Inconsistent quality
193 * Not scalable
194
195 **What Replaced It:**
196
197 * Better quality gates
198 * Sampling audits for system improvement
199 * Transparent confidence scoring
200 * Risk-based warnings
201
202 == 3. Content States ==
203
204 === 3.1 Published ===
205
206 **Status**: Visible to all users
207
208 **Includes:**
209
210 * AI-generated analyses (default state after passing gates)
211 * User-contributed content
212 * Edited/improved content
213
214 **Quality Indicators** (displayed with content):
215
216 * **Confidence Score**: 0-100% (AI's confidence in analysis)
217 * **Source Quality Score**: 0-100% (based on source track record)
218 * **Controversy Flag**: If high dispute/edit activity
219 * **Completeness Score**: % of expected fields filled
220 * **Last Updated**: Date of most recent change
221 * **Edit Count**: Number of revisions
222 * **Review Status**: AI-generated / Enhanced by contributors
223
224 **Automatic Warnings:**
225
226 * Confidence < 60%: "Low confidence - use caution"
227 * Source quality < 40%: "Sources may be unreliable"
228 * High controversy: "Disputed - multiple interpretations exist"
229 * Medical/Legal/Safety domain: "Seek professional advice"
230
231 **User Needs served**: UN-1 (Trust score), UN-9 (Methodology transparency), ~~UN-15 (Evolution timeline - Deferred)~~, UN-16 (Review status)
232
233
234 === 3.2 Hidden ===
235
236 **Status**: Not visible to regular users (only to moderators)
237
238 **Reasons:**
239
240 * Spam or advertising
241 * Personal attacks or harassment
242 * Illegal content
243 * Privacy violations
244 * Deliberate misinformation (verified)
245 * Abuse or harmful content
246
247 **Process:**
248
249 * Automated detection flags for moderator review
250 * Moderator confirms and hides
251 * Original author notified with reason
252 * Can appeal to board if disputes moderator decision
253
254 **Note**: Content is hidden, not deleted (for audit trail)
255
256
257 == 4. Role Evolution Path ==
258
259 **Contributor Journey:**
260
261 1. **Visitor** – Explores platform, reads documentation, may raise questions
262 2. **New Contributor** – Submits first improvements (typo fixes, small clarifications, new issues)
263 3. **Contributor** – Contributes regularly and follows project conventions
264 4. **Trusted Contributor** – Has a track record of high-quality work and reliable judgment
265 5. **Auditor** – Participates in sampling audits (pattern analysis)
266 6. **Moderator** – Focuses on behavior, tone, and conflict moderation (not content quality)
267 7. **Domain Expert** (optional) – Offers domain expertise without changing governance authority
268
269 **Key Principle**: Low barrier to entry, transparent criteria for advancement, clear separation between content quality (automated) and behavior moderation (human).
270
271
272 == 5. Principles ==
273
274 * **Low barrier to entry** for new contributors
275 * **Transparent criteria** for gaining and losing responsibilities
276 * **Clear separation** between content quality (automated) and behavioral moderation (human)
277 * **Documented processes** for escalation and appeal
278 * **No gatekeeping** for content publication
279 * **Immediate application** of contributions (Wikipedia model)
280
281 == 6. Related Pages ==
282
283 * [[AKEL>>Archive.FactHarbor 2026\.01\.20.Specification.AI Knowledge Extraction Layer (AKEL).WebHome]] - AI system
284 * [[Workflows>>Archive.FactHarbor 2026\.01\.20.Specification.Workflows.WebHome]] - Process workflows
285 * [[Architecture>>Archive.FactHarbor 2026\.01\.20.Specification.Architecture.WebHome]] - System architecture
286 * [[Decision Processes>>FactHarbor.Organisation.Decision-Processes.WebHome]] - Governance
287 * [[Requirements>>Archive.FactHarbor 2026\.01\.20.Specification.Requirements.WebHome]] - All requirements
288
289 **V0.9.70 CHANGES:**
290
291 **REMOVED:**
292 - All references to "Mode 3" publication
293 - "Contributors validate quality gates"
294 - "Trusted Contributors validate outputs"
295 - "Moderators finalize publication"
296
297 **CLARIFIED:**
298 - Contributors suggest system improvements (not approve outputs)
299 - Moderators handle abuse only (not content quality)
300 - Trusted Contributors provide consultation (not validation)
301 - Domain experts provide opinions (not approval)
302 - Only 2 publication modes (AI-Generated / Draft-Only)
303
304 **ADDED:**
305 - Clear role boundaries
306 - Explicit "cannot" lists for each role
307 - Publication mode details
308 - Content state specifications