Skip to Content

Wiki source code of Roles & Publication Modes

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