Recommendation of the Council on Artificial Intelligence (5/2/2024)

• Authority: OECD
• Jurisdiction: Multinational
• Collections: Editors' Picks
• Most recent activity: Enacted
• Most recent activity date: 2024-05-02

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• Unfair discrimination and misrepresentation
• Compromise of privacy
• AI system security vulnerabilities
• Overreliance and unsafe use
• Loss of human agency and autonomy
• Increased inequality and decline in employment
• Governance failure
• Lack of capability or robustness
• Lack of transparency or interpretability

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• Exposure to toxic content
• Unequal performance across groups
• False or misleading information
• Pollution of information ecosystem
• Disinformation, surveillance, and influence at scale
• Cyberattacks, weapon development or use
• Fraud, scams, and targeted manipulation
• Power centralization and unfair distribution
• Economic and cultural devaluation of human effort
• Environmental harm
• AI pursuing its own goals

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has good coverage of approximately 10-12 subdomains, with strong focus on discrimination and fairness (1.1, 1.3), privacy (2.1), misinformation (3.1, 3.2), malicious use (4.1, 4.2, 4.3), human agency (5.1, 5.2), socioeconomic impacts (6.1, 6.2, 6.3), governance (6.5), and system robustness (7.3, 7.4). Coverage is concentrated in fairness, transparency, accountability, and socioeconomic transformation domains.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Scientific Research & Development Services
• Public Administration (excl. National Security)

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the sector are described.

• Agriculture, Mining, Construction & Manufacturing
• Trade, Transportation & Utilities
• Information
• Finance & Insurance
• Professional & Technical Services
• Management, Administrative & Support Services
• Educational Services
• Health Care & Social Assistance
• Arts, Entertainment & Recreation
• Accommodation, Food & Other Services

All other sectors had no coverage in this document.

Sectors Summary

This is a cross-sectoral governance framework that applies broadly to all sectors where AI systems are developed, deployed, or used. The document does not limit its scope to specific industries but rather establishes principles and recommendations for trustworthy AI applicable across all economic sectors. Special attention is given to workplace transformation and labor markets across sectors.

Legislative Status

Soft Law

Legislative Status Summary

This is a non-binding international recommendation from the OECD that uses voluntary language throughout and relies on adherence rather than legal enforcement. The document explicitly 'RECOMMENDS' and 'INVITES' rather than mandates, with no penalties or formal enforcement mechanisms.

Actors

Proposer

• Entity Types: AI_Governance_Actor
• Entities: OECD; Digital Policy Committee
• Summary: The document is proposed by the OECD's Digital Policy Committee, as stated in the opening and closing sections. The OECD Council formally adopts the recommendation based on this proposal.

Target

• Entity Types: AI_Developer, AI_Deployer, AI_Governance_Actor
• Entities: Members and non-Members adhering to this Recommendation (Adherents); AI actors; Governments; stakeholders
• Summary: The document targets multiple entity types: governments (Adherents) who should implement national policies, AI actors (developers and deployers) who should follow the principles, and all stakeholders involved in AI systems. The principles apply to those who play active roles in the AI system lifecycle.

Enforcer

• Entity Types: AI_Governance_Actor
• Entities: Governments; OECD Digital Policy Committee
• Summary: Governments are responsible for implementing the recommendations through national policies and regulatory frameworks. The OECD Digital Policy Committee monitors implementation but does not enforce in a traditional regulatory sense, as this is a soft law instrument.

Monitor

• Entity Types: AI_Governance_Actor
• Entities: OECD Digital Policy Committee; Working Party on AI Governance
• Summary: The OECD Digital Policy Committee, through its Working Party on AI Governance, is explicitly tasked with monitoring implementation, developing measurement frameworks, providing forums for information exchange, and reporting to Council on implementation and continued relevance.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Collect and Process Data
• Build and Use Model
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document explicitly addresses all stages of the AI lifecycle, with particular emphasis on the entire lifecycle approach. It defines the AI system lifecycle comprehensively and requires principles and risk management to be applied throughout all phases.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• AI Systems
• Open Weight / Open Source

Technical Scope Summary

The document explicitly defines and addresses AI systems comprehensively, including their lifecycle, capabilities, and limitations. It does not use specific technical terms like 'frontier AI', 'GPAI', 'foundation models', or 'generative AI', nor does it specify compute thresholds. The focus is on AI systems broadly defined as machine-based systems that generate outputs influencing physical or virtual environments.