Global AI Governance Landscape

The global AI governance landscape involves multiple intersecting institutions, treaties, and standards bodies. Navigating it effectively matters for organisations building or procuring AI.

International standards

ISO/IEC 42001 (AI Management Systems) provides a certification framework for AI governance analogous to ISO 27001 for information security. NIST AI Risk Management Framework (AI RMF) from the US National Institute of Standards and Technology provides a comprehensive voluntary framework increasingly referenced in contracts and regulation globally.

Multi-lateral coordination

The OECD AI Principles (2019) established internationally agreed principles for trustworthy AI, signed by 46 countries including New Zealand and Australia. The Bletchley Declaration (2023) and subsequent AI Safety Summits represent emerging multilateral coordination on frontier AI safety. The UN's AI Advisory Body has released recommendations on international AI governance.

Industry self-regulation

Major AI providers have committed to voluntary safety measures including pre-deployment evaluation, red-teaming, and incident sharing. The effectiveness of voluntary commitments without binding enforcement is a subject of active policy debate.

Frontier AI safety

Separate from the regulation of current AI applications, significant resources are being devoted to the safety of future, more capable AI systems. Anthropic, OpenAI, DeepMind, and others have safety research teams. Government AI safety institutes (UK, US, Singapore, others) are conducting evaluations of frontier models. New Zealand participates in international AI safety discussions.

AI governance is fundamentally a coordination problem — getting many organisations and governments to agree on rules for technology that is borderless. The institutional machinery for this coordination is being built in real time.