Wiki source code of Automation Level Mermaid
Last modified by Robert Schaub on 2025/12/24 20:31
Hide last authors
| author | version | line-number | content |
|---|---|---|---|
 |
1.1 | 1 | {{mermaid}} |
| 2 | graph TD | ||
| 3 | L0[Level 0: AI-Publication POC | ||
| 4 | - AI-Generated Content Publication | ||
| 5 | - Quality Gates Active | ||
| 6 | - Contradiction Search Mandatory | ||
| 7 | - Risk Tier Classification | ||
| 8 | - Audit Sampling Started | ||
| 9 | - Recommendations: 30-50% Tier A, 10-20% B, 5-10% C] | ||
| 10 | |||
| 11 | L1[Level 1: Assisted Beta 0 | ||
| 12 | - AI drafts structures + publishes Tier C | ||
| 13 | - Humans audit samples | ||
| 14 | - Tier B mostly AI-published | ||
| 15 | - Tier A requires human review | ||
| 16 | - Feedback loop active | ||
| 17 | - Sampling rates adjust based on performance] | ||
| 18 | |||
| 19 | L2[Level 2: Structured Release 1.0 | ||
| 20 | - AI produces near-complete outputs | ||
| 21 | - Most Tier B AI-published | ||
| 22 | - Tier A human-reviewed for publication | ||
| 23 | - Mature audit system | ||
| 24 | - Lower sampling rates as confidence increases] | ||
| 25 | |||
| 26 | L3[Level 3: Distributed Intelligence Future | ||
| 27 | - Federated contradiction detection | ||
| 28 | - Cross-node audit sharing | ||
| 29 | - Advanced bubble detection | ||
| 30 | - Tier A still human-reviewed | ||
| 31 | - Strategic audits only] | ||
| 32 | |||
| 33 | L0 --> L1 | ||
| 34 | L1 --> L2 | ||
| 35 | L2 --> L3 | ||
| 36 | |||
| 37 | subgraph "Constant Across All Levels" | ||
| 38 | HumanAuth[Humans retain final authority | ||
| 39 | Quality gates mandatory | ||
| 40 | Tier A requires human review for Mode 3 | ||
| 41 | Audit system active | ||
| 42 | All numbers are recommendations] | ||
| 43 | end | ||
| 44 | |||
| 45 | classDef current fill:#e3f2fd,stroke:#1976d2,stroke-width:3px | ||
| 46 | classDef future fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px | ||
| 47 | |||
| 48 | class L0 current | ||
| 49 | class L3 future | ||
| 50 | {{/mermaid}} |