Skip to content

OECD-AI-PRINCIPLES-MAPPING.md

Latest commit

 

History

History
75 lines (56 loc) · 4.68 KB

OECD-AI-PRINCIPLES-MAPPING.md

File metadata and controls

75 lines (56 loc) · 4.68 KB
Raw

OECD AI Principles Mapping

Framework: OECD Recommendation on Artificial Intelligence (2019, updated 2024) Version: 0.1 Since: v0.11.3

This document maps the OECD AI Principles to PEAC Protocol capabilities.

Framework Overview

The OECD AI Principles provide values-based principles and policy recommendations for trustworthy AI. They are adopted by 46 countries and serve as the basis for many national AI strategies. PEAC provides verifiable evidence infrastructure aligned with these principles.

Principle Mapping

Principle 1: Inclusive Growth, Sustainable Development, and Well-Being

Aspect PEAC Mechanism
Benefit tracking Receipt evidence of AI-assisted interactions
Impact measurement Interaction evidence with structured outcomes
Stakeholder inclusion Open protocol (vendor-neutral, interoperable)

Principle 2: Human-Centred Values and Fairness

Aspect PEAC Mechanism Package
Privacy protection Hash-first evidence (SHA-256 digests, no raw PII) @peac/schema
Non-discrimination evidence Control chain recording decision reasons @peac/control
Human override Control action with manual_review trigger @peac/schema
Consent recording Purpose declaration + content signals @peac/protocol

Principle 3: Transparency and Explainability

Aspect PEAC Mechanism Package
Disclosure of AI use Agent identity in receipt sub claim @peac/kernel
Decision explanation Control chain with reasons per step @peac/control
Data provenance Content signals source precedence @peac/mappings-content-signals
Audit trail Signed receipts with offline verification @peac/protocol

Principle 4: Robustness, Security, and Safety

Aspect PEAC Mechanism Package
Cryptographic security EdDSA signatures, key rotation lifecycle @peac/crypto
SSRF protection URL validation, private IP blocking @peac/protocol
Replay resistance Unique jti per receipt, temporal bounds @peac/protocol
Key compromise response Emergency revocation, revoked_keys in issuer config @peac/protocol
Risk monitoring Risk signal observations ZT Profile Pack

Principle 5: Accountability

Aspect PEAC Mechanism Package
Actor identification ActorBinding with 8 proof types @peac/schema
Action attribution Interaction evidence per tool invocation @peac/schema
Evidence preservation Dispute bundles with reconciliation @peac/audit, @peac/cli
Deterministic reporting Reconcile CLI --format json output @peac/cli
Identity completeness MVIS enforcement (5 required fields) @peac/schema

OECD Policy Recommendations

Recommendation PEAC Alignment
Investing in AI R&D Open-source protocol; permissive licensing
Fostering a digital ecosystem Interoperable evidence carriers (MCP, A2A, HTTP, x402)
Providing an enabling policy environment Policy discovery via peac.txt and peac-issuer.json
Building human capacity Plugin Pack for developer tools
International cooperation Vendor-neutral, standards-aligned (RFC 8032, RFC 9711)

References