As artificial intelligence systems grow more capable and widely deployed, global standards for AI models have emerged as a critical mechanism for aligning national regulations, establishing shared technical norms, and supporting the safe governance of both general-purpose and frontier systems. The current landscape spans overlapping soft-law frameworks, technical standards, and binding regulations developed across institutions including the OECD, UNESCO, G7, EU, UN, ISO, NIST, and various industry consortia.
The OECD AI Principles, first adopted in 2019 and updated in 2024, represent the first intergovernmental AI standard and have been endorsed by more than 45 countries, making them the most widely adopted international AI governance benchmark. These principles promote innovative, trustworthy AI that respects human rights, democratic values, and the rule of law, while calling for multi-stakeholder global technical standards to support AI interoperability.
Endorsed by over 45 countries, the OECD AI Principles remain the world’s most widely adopted international AI governance benchmark.
The UNESCO Recommendation on the Ethics of AI, adopted in 2021 by all 193 UNESCO member states, further established globally shared values and policy areas for ethical AI development and use.
At the political level, the UN General Assembly passed an AI resolution in March 2024 representing the first global consensus on safe, secure, and trustworthy AI, encouraging international cooperation around standards and governance mechanisms. The G7 Hiroshima AI Process, launched in May 2023, issued guiding principles and a voluntary code of conduct for organizations developing advanced AI systems, covering commitments to safety testing, information-sharing, and content labeling.
Both frameworks increasingly reference international technical standards and propose coordinated mechanisms, such as an AI standards exchange, to catalogue and harmonize applicable norms across jurisdictions.
On the technical side, ISO/IEC 42001:2023 defines the first certifiable management system standard for AI, enabling organizations to build structured governance, risk, and compliance processes. ISO/IEC 23894:2023 provides lifecycle-based guidance for AI-specific risk identification, assessment, and mitigation across development, deployment, and operation.
ISO/IEC 42005:2025 introduces an AI system impact assessment standard to evaluate organizational and societal impacts throughout the AI lifecycle, while ISO/IEC 5259-2 addresses data quality requirements for reliable training, validation, and operation of AI models.
The EU AI Act reinforces this technical standards ecosystem by introducing an extensive, risk-based regulatory framework that references and depends on technical standards for conformity assessment and compliance. National and regional governance initiatives are increasingly aligning with global reference frameworks such as the OECD AI Principles, NIST AI Risk Management Framework, and ISO/IEC AI standards to reduce regulatory fragmentation.
The convergence of intergovernmental principles, voluntary codes of conduct, and certifiable technical standards reflects a broad recognition that no single jurisdiction can effectively govern frontier AI systems in isolation, and that coordinated global frameworks are essential to managing the risks and opportunities that advanced AI presents. Notably, 87% of organizations plan to implement AI governance frameworks by 2026, underscoring the accelerating institutional momentum behind coordinated global standards.
