---
type: CKG Bundle
title: AI Governance
description: "'The AI told me to' is not a defense — immutable reasoning trails, provenance, and audit-ready decisions."
tags: [governance, risk, audit, compliance, ckg]
timestamp: 2026-06-18T00:00:00Z
ckg:
  id: ai-governance
  version: 1.0.0
  source: "Graphify.md positioning + EU AI Act / NIST AI RMF"
---
# AI Governance — CKG

```csv
ConceptID,ConceptLabel,Dependencies,TaxonomyID
1,AI decision,,FOUND
2,Accountability,1,GOV
3,"The AI told me to" (not a defense),1,GOV
4,Auditability,,GOV
5,Reasoning trail,4,GOV
6,Immutable log,5,GOV
7,Provenance,,FIX
8,Cited source,7,FIX
9,Compressed Knowledge Graph,,CKG
10,Deterministic traversal,9,CKG
11,Human attestation (L2),9,CKG
12,Conformance profile,11,CKG
13,EU AI Act / NIST AI RMF,,REG
14,Audit-ready decision,6|8|11,GOV
```

## Edges
1→2 · 1→3 · 4→5 · 5→6 · 7→8 · 9→10 · 9→11 · 11→12 · 6→14 · 8→14 · 11→14 · 13→14

> "The AI told me to" won't hold up. A CKG makes every answer a traversal of declared, sourced edges — cited, logged, and signed (L2) by a named authority. That's a decision you can defend. *Built by Graphify.md.*
