OECD AI Principles (2019, revised 2024)
The OECD AI Principles (2019, revised 2024) are the first intergovernmental AI standard, adopted by 47 countries. They articulate five values — inclusive growth, human rights and democratic values, transparency and explainability, robustness/security/safety, and accountability. Ledgix evidences each principle via the ledger, policy snapshots, and cryptographic custody.
Status: Full — every control resolves to an artifact Ledgix produces today.
Scope
The OECD AI Principles apply voluntarily to AI actors in adherent jurisdictions (47 countries as of 2024, including all G7 and G20 members). They form the basis for several downstream AI frameworks and are explicitly referenced by the EU AI Act recitals and by Brazil's PL 2338. Coverage spans inclusive growth and well-being, human rights and democratic values, transparency and explainability, robustness/security/safety, and accountability.
Controls covered
| Field | Type | Required | Description |
|---|---|---|---|
| OECD-AI-1.1 | policy_snapshots / events_jsonl | Inclusive growth, sustainable development and well-being | Policies articulate values and boundaries; decisions linked to governing policies and human principals. |
| OECD-AI-1.2 | events_jsonl | Human rights and democratic values (fairness and privacy) | Each decision's rationale referencing policy chunks demonstrates values-aware evaluation. |
| OECD-AI-1.3 | events_jsonl | Transparency and explainability | Per-decision reason, citations, evidence chunks, and confidence. |
| OECD-AI-1.4 | checkpoint_chain / key_history / signatures | Robustness, security and safety | Continuous Merkle integrity, signing-key custody, and export signature. |
| OECD-AI-1.5 | events_jsonl / proof_index | Accountability | Ed25519 receipts tying actions to AI system instances plus Merkle inclusion proofs. |
Evidence types referenced
- policy_snapshots — policies articulating the values and boundaries guiding AI actions.
- events_jsonl — per-decision reasoning, citations, evidence chunks, and accountability metadata.
- checkpoint_chain — continuous Merkle-chained integrity record.
- key_history — signing-key lifecycle for key custody controls.
- signatures — export signature for independent verification.
- proof_index — Merkle inclusion proofs for independent re-verification.
Known gaps (if any)
None — every control resolves to an artifact Ledgix produces today.
Audit pack workflow
Export an evidence ZIP for this framework from the admin console's Evidence Exports panel by selecting OECD AI Principles (2019, revised 2024) and a time window. Each control's evidence_locators[] in the included framework_mapping.json points to the corresponding file in the ZIP.
References
- Framework mapping JSON:
vault/internal/compliance/frameworks/oecd_ai_principles.json - Canonical source: OECD AI Principles — oecd.ai