Stanford CodeX AILCCP: 37 Principles Wired to Controls, Standards, and Risks
Eran Kahana’s Stanford CodeX post ships AILCCP — a knowledge graph of 37 principles, 48 controls, 43 standards, 10 life-cycle phases, 18 risks, 500+ links, plus Explorer at ailccp.replit.app. The trade: make ‘trustworthy’ operational and auditable across instruments that don’t interoperate (ISO 42001, NIST AI RMF, IEEE, EU AI Act).
AI governance fails less from missing slogans than from missing maps. AILCCP’s frame: stitch OECD/UNESCO/G7-style principles to controls, ISO/IEEE/NIST standards, life-cycle owners, and rated risks so “ethical” stops being undefined wallpaper.
Eran Kahana on the Stanford CodeX blog (April 5) detailed the AI Life Cycle Core Principles (AILCCP) framework: a navigable knowledge graph with 37 principles, 48 controls, 43 international standards, 10 life-cycle phases (scoping → decommissioning), 18 risks, and 500+ explicit cross-references — plus ownership, metrics, and evidence artifacts. Interactive AILCCP Explorer: ailccp.replit.app.
Graph contents (author counts)#
| Layer | Count / notes |
|---|---|
| Principles | 37 — defs, objectives, questions, controls, evidence, stakeholders; 15 categories / 10 pillars |
| Controls | 48 — 187 control↔principle links |
| Standards | 43 IEEE/ISO/IEC/NIST — 215 standard↔principle links (29 principles) |
| Life cycle | 10 phases — owners (Product, Legal, ML Eng…), artifacts, metrics; 84 phase↔principle links |
| Risks | 18 — 7 Very High / 8 High / 3 Medium; 23 risk↔standard links; “enabling risks” (e.g. transparency gaps) |
Problem diagnosis: ISO/IEC 42001, NIST AI RMF, IEEE, EU AI Act don’t interoperate cleanly — orgs reconcile without principle→control→phase→risk maps.
Claims vs checks#
Component counts and link totals are from the April 5 CodeX publication. Usefulness depends on org adoption — not proven by the blog post. Explorer is a Replit app; treat as author tooling, not a standards body adoption.
Limits#
- Academic/practitioner framework — not law.
- Crosswalks can over-simplify jurisdictional conflicts.
- Maintenance burden as standards revise.
Sources#
- Eran Kahana, Stanford CodeX: “Turning AI Governance Into Operational Infrastructure” (April 5, 2026).
- AILCCP Explorer.
- Framework details from the April 5 publication (counts and cross-reference totals).