International AI Governance Frameworks

1. Overview

International AI governance operates through a complex web of multilateral frameworks, each with different membership, legal status, and scope. Understanding this landscape is essential because national AI regulations are increasingly shaped by — and measured against — these international norms.

2. OECD AI Principles

2.1 Background

The OECD Recommendation on Artificial Intelligence (May 2019, updated May 2024) was the first intergovernmental standard on AI. It has been adopted by 46 countries and has influenced virtually every national AI strategy developed since.

2.2 The Five Principles

2.3 2024 Update

The 2024 revision updated the principles to address developments since 2019:

• Generative AI: Expanded scope to explicitly cover foundation models and generative AI
• Interoperability: Emphasis on interoperability of AI governance across jurisdictions
• Environmental impact: New attention to AI’s energy consumption and environmental footprint
• Information integrity: New principle area on AI and misinformation/disinformation

2.4 OECD AI Policy Observatory

The OECD maintains the AI Policy Observatory (OECD.AI), the most comprehensive database of national AI policies worldwide, tracking 800+ policy initiatives across 70+ countries.

3. UNESCO Recommendation on the Ethics of AI

3.1 Background

Adopted by all 193 UNESCO member states in November 2021, this is the first global normative instrument on AI ethics. While non-binding, its universal adoption gives it significant normative weight.

3.2 Values & Principles

3.3 Readiness Assessment Methodology

UNESCO developed a Readiness Assessment Methodology (RAM) to help countries evaluate their preparedness for ethical AI governance. As of 2025, over 50 countries have completed or are conducting RAM assessments, identifying gaps in legal frameworks, institutional capacity, and technical infrastructure.

4. Council of Europe Framework Convention on AI

4.1 Historic Significance

4.2 Key Provisions

• Scope: Applies to AI systems throughout their lifecycle, covering public and private sector use
• Human rights protection: Parties must ensure AI respects human rights as protected by the European Convention on Human Rights and other international instruments
• Transparency: Persons interacting with AI must be informed; ability to understand AI decisions affecting rights
• Accountability: Effective remedies for human rights violations caused by AI
• Risk management: Risk-based approach to AI governance with assessments proportionate to risk
• National security exception: Parties may exclude national security activities, but must ensure such activities respect international law

4.3 Signatories

As of early 2026, the treaty has been signed by the EU, US, UK, Canada, Japan, Israel, and multiple CoE member states. Ratification processes are underway in several countries.

5. G7 Hiroshima AI Process

5.1 Background

Launched at the G7 Summit in Hiroshima (May 2023), the Hiroshima AI Process focused specifically on governance of advanced AI systems, particularly foundation models and generative AI.

5.2 Outputs

5.3 The 11 Guiding Principles

1. Take appropriate measures throughout the development of advanced AI systems
2. Identify and mitigate risks across the AI lifecycle
3. Report advanced AI capabilities, limitations, and domains of appropriate/inappropriate use
4. Invest in responsible AI development
5. Develop and implement robust information security measures
6. Develop and deploy reliable content authentication and provenance mechanisms
7. Prioritize research to mitigate societal, safety, and security risks
8. Develop and deploy advanced AI systems for identified societal challenges
9. Invest in developing standards, tools, and practices for responsible AI
10. Support the development and adoption of international AI standards
11. Implement appropriate data governance and protection measures

6. Global Partnership on AI (GPAI)

6.1 Structure

GPAI was launched in June 2020 by 15 founding members and has grown to 29 members. In December 2024, GPAI merged with the OECD, becoming the OECD’s AI governance hub.

6.2 Working Groups

7. United Nations Activities

7.1 UN Secretary-General’s High-Level Advisory Body on AI

Established in 2023, the 39-member advisory body published its interim report (“Governing AI for Humanity”) in December 2023 and final report in September 2024:

• Key recommendation: New international AI governance institution(s) under UN auspices
• Proposed functions: International scientific panel on AI; policy dialogue forum; standards coordination; capacity building
• Global Digital Compact (2024): AI governance included as key component of the GDC adopted at the Summit of the Future

7.2 UN Agency Activities

8. Other Multilateral Initiatives

9. Comparative Analysis

10. References & Resources

OECD

• OECD Recommendation on Artificial Intelligence (2024 update)
• OECD AI Policy Observatory (OECD.AI) — Database of 800+ AI policies worldwide

UNESCO

• UNESCO Recommendation on the Ethics of AI (2021)
• UNESCO Readiness Assessment Methodology

Council of Europe

• CoE Framework Convention on AI — Treaty text and status

G7 & GPAI

• G7 Hiroshima AI Process — Principles and Code of Conduct
• Global Partnership on AI (GPAI)

United Nations

• UN High-Level Advisory Body on AI
• Global Digital Compact (2024)