Skip to Content

Changes for page Workflows

Last modified by Robert Schaub on 2025/12/24 20:34
From version 7.2
edited by Robert Schaub
on 2025/12/16 20:28
Change comment: Renamed back-links.
To version 6.1
edited by Robert Schaub
on 2025/12/14 18:59
Change comment: Imported from XAR

Summary

Details

Page properties
Content
... ... @@ -21,7 +21,6 @@
21 21  **Actor**: Contributor or AKEL
22 22  
23 23  **Actions**:
24 -
25 25  * Submit claim text
26 26  * Provide initial sources (optional for human contributors, mandatory for AKEL)
27 27  * System assigns initial AuthorType (Human or AI)
... ... @@ -31,7 +31,6 @@
31 31  === Step 2: AKEL Processing ===
32 32  
33 33  **Automated Steps**:
34 -
35 35  1. Claim extraction and normalization
36 36  2. Classification (domain, type, evaluability)
37 37  3. Risk tier assignment (A/B/C suggested)
... ... @@ -45,7 +45,6 @@
45 45  === Step 3: Quality Gate Checkpoint ===
46 46  
47 47  **Gates Evaluated**:
48 -
49 49  * Source quality
50 50  * Contradiction search completion
51 51  * Uncertainty quantification
... ... @@ -52,7 +52,6 @@
52 52  * Structural integrity
53 53  
54 54  **Outcomes**:
55 -
56 56  * **All gates pass** → Proceed to Mode 2 publication (if Tier B or C)
57 57  * **Any gate fails** → Mode 1 (Draft), flag for human review
58 58  * **Tier A** → Mode 2 with warnings + auto-escalate to expert queue
... ... @@ -60,19 +60,16 @@
60 60  === Step 4: Publication (Risk-Tier Dependent) ===
61 61  
62 62  **Tier C (Low Risk)**:
63 -
64 64  * **Direct to Mode 2**: AI-generated, public, clearly labeled
65 65  * User can request human review
66 66  * Sampling audit applies
67 67  
68 68  **Tier B (Medium Risk)**:
69 -
70 70  * **Direct to Mode 2**: AI-generated, public, clearly labeled
71 71  * Higher audit sampling rate
72 72  * High-engagement content may auto-escalate
73 73  
74 74  **Tier A (High Risk)**:
75 -
76 76  * **Mode 2 with warnings**: AI-generated, public, prominent disclaimers
77 77  * **Auto-escalated** to expert review queue
78 78  * User warnings displayed
... ... @@ -81,7 +81,6 @@
81 81  === Step 5: Human Review (Optional for B/C, Escalated for A) ===
82 82  
83 83  **Triggers**:
84 -
85 85  * User requests review
86 86  * Audit flags issues
87 87  * High engagement (Tier B)
... ... @@ -88,7 +88,6 @@
88 88  * Automatic (Tier A)
89 89  
90 90  **Process**:
91 -
92 92  1. Reviewer/Expert examines claim
93 93  2. Validates quality gates
94 94  3. Checks contradiction search results
... ... @@ -96,7 +96,6 @@
96 96  5. Decision: Approve, Request Changes, or Reject
97 97  
98 98  **Outcomes**:
99 -
100 100  * **Approved** → Mode 3 (Human-Reviewed)
101 101  * **Changes Requested** → Back to contributor or AKEL for revision
102 102  * **Rejected** → Rejected status with reasoning
... ... @@ -108,13 +108,11 @@
108 108  === Step 1: Scenario Generation ===
109 109  
110 110  **Automated (AKEL)**:
111 -
112 112  * Generate scenarios for claim
113 113  * Define boundaries, assumptions, context
114 114  * Identify evaluation methods
115 115  
116 116  **Manual (Expert/Reviewer)**:
117 -
118 118  * Create custom scenarios
119 119  * Refine AKEL-generated scenarios
120 120  * Add domain-specific nuances
... ... @@ -122,7 +122,6 @@
122 122  === Step 2: Scenario Validation ===
123 123  
124 124  **Quality Checks**:
125 -
126 126  * Completeness (definitions, boundaries, assumptions clear)
127 127  * Relevance to claim
128 128  * Evaluability
... ... @@ -129,7 +129,6 @@
129 129  * No circular logic
130 130  
131 131  **Risk Tier Assignment**:
132 -
133 133  * Inherits from parent claim
134 134  * Can be overridden by expert if scenario increases/decreases risk
135 135  
... ... @@ -136,12 +136,10 @@
136 136  === Step 3: Scenario Publication ===
137 137  
138 138  **Mode 2 (AI-Generated)**:
139 -
140 140  * Tier B/C scenarios can publish immediately
141 141  * Subject to sampling audits
142 142  
143 143  **Mode 1 (Draft)**:
144 -
145 145  * Tier A scenarios default to draft
146 146  * Require expert validation for Mode 2 or Mode 3
147 147  
... ... @@ -152,7 +152,6 @@
152 152  === Step 1: Evidence Search & Retrieval ===
153 153  
154 154  **AKEL Actions**:
155 -
156 156  * Search academic databases, reputable media
157 157  * **Mandatory contradiction search** (counter-evidence, reservations)
158 158  * Extract metadata (author, date, publication, methodology)
... ... @@ -159,7 +159,6 @@
159 159  * Assess source reliability
160 160  
161 161  **Quality Requirements**:
162 -
163 163  * Primary sources preferred
164 164  * Diverse perspectives included
165 165  * Echo chambers flagged
... ... @@ -168,7 +168,6 @@
168 168  === Step 2: Evidence Summarization ===
169 169  
170 170  **AKEL Generates**:
171 -
172 172  * Summary of evidence
173 173  * Relevance assessment
174 174  * Reliability score
... ... @@ -180,7 +180,6 @@
180 180  === Step 3: Evidence Review ===
181 181  
182 182  **Reviewer/Expert Validates**:
183 -
184 184  * Accuracy of summaries
185 185  * Appropriateness of sources
186 186  * Completeness of contradiction search
... ... @@ -187,7 +187,6 @@
187 187  * Reliability assessments
188 188  
189 189  **Outcomes**:
190 -
191 191  * **Mode 2**: Evidence summaries published as AI-generated
192 192  * **Mode 3**: After human validation
193 193  * **Mode 1**: Failed quality checks or pending expert review
... ... @@ -199,7 +199,6 @@
199 199  === Step 1: Verdict Computation ===
200 200  
201 201  **AKEL Computes**:
202 -
203 203  * Verdict across scenarios
204 204  * Confidence scores
205 205  * Uncertainty quantification
... ... @@ -207,7 +207,6 @@
207 207  * Limitations
208 208  
209 209  **Inputs**:
210 -
211 211  * Claim text
212 212  * Scenario definitions
213 213  * Evidence assessments
... ... @@ -216,13 +216,11 @@
216 216  === Step 2: Verdict Validation ===
217 217  
218 218  **Quality Gates**:
219 -
220 220  * All four gates apply (source, contradiction, uncertainty, structure)
221 221  * Reasoning chain must be traceable
222 222  * Assumptions must be explicit
223 223  
224 224  **Risk Tier Check**:
225 -
226 226  * Tier A: Always requires expert validation for Mode 3
227 227  * Tier B: Mode 2 allowed, audit sampling
228 228  * Tier C: Mode 2 default
... ... @@ -230,7 +230,6 @@
230 230  === Step 3: Verdict Publication ===
231 231  
232 232  **Mode 2 (AI-Generated Verdict)**:
233 -
234 234  * Clear labeling with confidence scores
235 235  * Uncertainty disclosure
236 236  * Links to reasoning trail
... ... @@ -237,7 +237,6 @@
237 237  * User can request expert review
238 238  
239 239  **Mode 3 (Expert-Validated Verdict)**:
240 -
241 241  * Human reviewer/expert stamp
242 242  * Additional commentary (optional)
243 243  * Highest trust level
... ... @@ -249,7 +249,6 @@
249 249  === Step 1: Audit Sampling Selection ===
250 250  
251 251  **Stratified Sampling**:
252 -
253 253  * Risk tier priority (A > B > C)
254 254  * Low confidence scores
255 255  * High traffic content
... ... @@ -257,7 +257,6 @@
257 257  * User flags
258 258  
259 259  **Sampling Rates** (Recommendations):
260 -
261 261  * Tier A: 30-50%
262 262  * Tier B: 10-20%
263 263  * Tier C: 5-10%
... ... @@ -265,7 +265,6 @@
265 265  === Step 2: Audit Execution ===
266 266  
267 267  **Auditor Actions**:
268 -
269 269  1. Review sampled AI-generated content
270 270  2. Validate quality gates were properly applied
271 271  3. Check contradiction search completeness
... ... @@ -273,7 +273,6 @@
273 273  5. Identify errors or hallucinations
274 274  
275 275  **Audit Outcome**:
276 -
277 277  * **Pass**: Content remains in Mode 2, logged as validated
278 278  * **Fail**: Content flagged for review, system improvement triggered
279 279  
... ... @@ -280,7 +280,6 @@
280 280  === Step 3: Feedback Loop ===
281 281  
282 282  **System Improvements**:
283 -
284 284  * Failed audits analyzed for patterns
285 285  * AKEL parameters adjusted
286 286  * Quality gates refined
... ... @@ -287,7 +287,6 @@
287 287  * Risk tier assignments recalibrated
288 288  
289 289  **Transparency**:
290 -
291 291  * Audit statistics published periodically
292 292  * Patterns shared with community
293 293  * System improvements documented
... ... @@ -299,7 +299,6 @@
299 299  === Mode 1 → Mode 2 ===
300 300  
301 301  **Requirements**:
302 -
303 303  * All quality gates pass
304 304  * Risk tier B or C (or A with warnings)
305 305  * Contradiction search completed
... ... @@ -309,7 +309,6 @@
309 309  === Mode 2 → Mode 3 ===
310 310  
311 311  **Requirements**:
312 -
313 313  * Human reviewer/expert validation
314 314  * Quality standards confirmed
315 315  * For Tier A: Expert approval required
... ... @@ -320,13 +320,11 @@
320 320  === Mode 3 → Mode 1 (Demotion) ===
321 321  
322 322  **Rare - Only if**:
323 -
324 324  * New evidence contradicts verdict
325 325  * Error discovered in reasoning
326 326  * Source retraction
327 327  
328 328  **Process**:
329 -
330 330  1. Content flagged for re-evaluation
331 331  2. Moved to draft (Mode 1)
332 332  3. Re-processed through workflow
... ... @@ -339,12 +339,10 @@
339 339  === On Mode 1 (Draft) Content ===
340 340  
341 341  **Contributors**:
342 -
343 343  * Edit their own drafts
344 344  * Submit for review
345 345  
346 346  **Reviewers/Experts**:
347 -
348 348  * View and comment
349 349  * Request changes
350 350  * Approve for Mode 2 or Mode 3
... ... @@ -352,7 +352,6 @@
352 352  === On Mode 2 (AI-Generated) Content ===
353 353  
354 354  **All Users**:
355 -
356 356  * Read and use content
357 357  * Request human review
358 358  * Flag for expert attention
... ... @@ -359,7 +359,6 @@
359 359  * Provide feedback
360 360  
361 361  **Reviewers/Experts**:
362 -
363 363  * Validate for Mode 3 transition
364 364  * Edit and refine
365 365  * Adjust risk tier if needed
... ... @@ -367,12 +367,10 @@
367 367  === On Mode 3 (Human-Reviewed) Content ===
368 368  
369 369  **All Users**:
370 -
371 371  * Read with highest confidence
372 372  * Still can flag if new evidence emerges
373 373  
374 374  **Reviewers/Experts**:
375 -
376 376  * Update if needed
377 377  * Trigger re-evaluation if new evidence
378 378  
... ... @@ -380,26 +380,10 @@
380 380  
381 381  == Diagram References ==
382 382  
383 -=== Claim and Scenario Lifecycle (Overview) ===
340 +{{include reference="FactHarbor.Organisation.Diagrams.Claim-Scenario-Lifecycle"/}}
384 384  
385 -{{include reference="FactHarbor.Archive.FactHarbor V0\.9\.23 Lost Data.Organisation.Diagrams.Claim and Scenario Lifecycle (Overview).WebHome"/}}
342 +{{include reference="FactHarbor.Specification.Diagrams.Manual vs Automated matrix.WebHome"/}}
386 386  
387 -=== Claim and Scenario Workflow ===
388 -
389 -{{include reference="Test.FactHarborV09.Specification.Diagrams.Claim and Scenario Workflow.WebHome"/}}
390 -
391 -=== Evidence and Verdict Workflow ===
392 -
393 -{{include reference="Test.FactHarborV09.Specification.Diagrams.Evidence and Verdict Workflow.WebHome"/}}
394 -
395 -=== Quality and Audit Workflow ===
396 -
397 -{{include reference="Test.FactHarborV09.Specification.Diagrams.Quality and Audit Workflow.WebHome"/}}
398 -
399 -
400 -
401 -{{include reference="Test.FactHarborV09.Specification.Diagrams.Manual vs Automated matrix.WebHome"/}}
402 -
403 403  ----
404 404  
405 405  == Related Pages ==
... ... @@ -408,3 +408,4 @@
408 408  * [[Automation>>FactHarbor.Specification.Automation.WebHome]]
409 409  * [[Requirements (Roles)>>FactHarbor.Specification.Requirements.WebHome]]
410 410  * [[Governance>>FactHarbor.Organisation.Governance]]
352 +