International AI Standards: ISO, OECD, and UN
International standards bodies produce the AI governance frameworks that work across borders without binding legal force. ISO/IEC 42001:2023 is the first international standard for AI management systems, and organizations can be certified against it. ISO/IEC 23894 covers AI risk management, and the Technical Report ISO/IEC TR 24028 gives informative guidance on AI trustworthiness.
The OECD AI Principles are the most influential international framework in AI policy. Many national AI strategies draw on them, and they informed both the EU AI Act and the NIST AI Risk Management Framework. UN bodies add the UNESCO Recommendation on the Ethics of AI and the newer UN Global Dialogue on AI Governance.
ISO/IEC 42001 certification helps organizations that manage compliance in several jurisdictions at once. It maps well to the governance duties in the EU AI Act and to the NIST framework's Govern function. In Europe, EN 18286 adds a quality management standard written for the AI Act itself.
Key themes
- 1.ISO/IEC 42001:2023, the certifiable AI management system standard
- 2.ISO/IEC 23894 for AI risk management
- 3.OECD AI Principles as the cross-border baseline
- 4.Fit with the EU AI Act and the NIST AI RMF
Regulatory frameworks and guidance(8)
ISO/IEC 42001:2023 - Artificial Intelligence Management System
ISO and IEC published ISO/IEC 42001:2023 in December 2023 as the first international AI management system standard. It sets requirements for establishing, maintaining, and improving an AI Management System. Organizations developing or using AI products and services can seek independent certification.
ISO/IEC 23894 AI Risk Management
ISO/IEC 23894 explains how to integrate AI risks into enterprise risk management. It adapts ISO 31000 risk terms and processes to traits specific to AI. These include unexpected behavior, reliance on data, hard-to-explain decisions, and complex effects on people and organizations.
ISO/IEC TR 24028:2020, Overview of Trustworthiness in Artificial Intelligence
ISO/IEC 24028:2020 explains AI trustworthiness concepts, characteristics, and threats. It provides technical and organizational approaches for assessing and improving trustworthiness throughout an AI system's life.
ISO/IEC 24029 Robustness of Neural Networks (TR 24029-1:2021 and 24029-2:2023)
ISO/IEC 24029 covers mathematical proof methods and practical assessments of artificial neural network robustness (how reliably a network performs when its inputs are unusual or manipulated). Its parts address deliberately misleading inputs, shifts in real-world data, and other failure modes relevant to trustworthy deployment.
OECD AI Principles
The OECD AI Principles were the first intergovernmental AI standard. They set five values-based principles and five government recommendations supporting trustworthy AI, human rights, and democratic values.
OECD Report: Governing with Artificial Intelligence
This OECD report examines government AI use across member and partner countries. Public-service improvements account for 57% of documented applications, while 45% support administrative decisions. Identified risks include biased data used to train systems, limited transparency, and overreliance on automated outputs.
UN Independent International Scientific Panel on AI: Preliminary Report and AI Agents Thematic Brief
The UN's Independent International Scientific Panel on AI released its preliminary report on AI opportunities, risks, and impacts on 1 July 2026. Its first thematic brief, on AI agents and the risk of losing human control, followed on 21 September 2026. The Panel provides scientific assessment and does not set rules.
UNESCO Recommendation on the Ethics of Artificial Intelligence: 2026 Implementation Tools Update
UNESCO launched three tools under its 2021 Recommendation on the Ethics of AI in September 2026. It presented them at its fourth Global Forum on the Ethics of AI in Riyadh. They are for policymakers: an updated Readiness Assessment Methodology (RAM 2.0), an analysis of RAM findings, and an AI environment toolkit.