Artificial intelligence (AI) is transforming the world at extraordinary speed. It is reshaping manufacturing, finance, healthcare, education, scientific research, and even the way governments make decisions. While AI is advancing at an unprecedented pace, one fundamental question has come to the fore: how to govern intelligence in an interconnected yet increasingly fragmented world?
The question came into sharper focus after Chinese President Xi Jinping delivered the opening address at the World Artificial Intelligence Conference (WAIC) in Shanghai. Much of the international commentary focused on China’s technological ambitions and the geopolitical implications of his speech. Equally significant, however, was the broader question it raised: if AI is becoming one of the defining technologies of the twenty-first century, what kind of international governance architecture will be needed to ensure that it will serve humanity as a whole?
The international debate on AI has evolved rapidly over the past few years. Initial discussions centred on use of resources and displacement of employees. Those concerns remain important, but they are no longer the central issue. The debate has moved from what AI can do to how AI should be governed. The real challenge is therefore no longer technological innovation; it is institutional adaptation.
The United States, the European Union, and China have each offered different approaches. The United States understandably places priority on maintaining technological leadership and innovation. The European Union has focused on protecting fundamental rights through risk-based regulation, transparency, accountability, and protection of privacy. China sees AI as a strategic engine for economic development and industrial upgrading, while also emphasizing security, ethical governance, and international cooperation.
These differences often dominate international headlines. Yet they may obscure a more important reality. Despite their different priorities, all three recognise that AI is becoming too powerful, too pervasive and too interconnected to be treated as an ordinary technology.
Throughout history, transformative technologies have repeatedly forced societies to redesign their institutions. Industrialisation transformed labour regulation. Global finance gave rise to international banking standards. Climate change has required entirely new forms of international cooperation. As institutional economist Douglas North observed, institutions evolve as changing economic conditions create incentives and uncertainties that existing rules can no longer manage effectively. AI may represent the latest example of this process.
Unlike previous technological revolutions, AI is not confined to a single sector. It is a general-purpose technology capable of influencing virtually every part of the economy and society simultaneously. It shapes productivity, labour markets, scientific discovery, financial systems, education, healthcare, public administration and national security. More importantly, it is beginning to reshape decision-making itself.
This is why I believe the international discussion needs a broader conceptual framework. AI governance should not be viewed simply as another area of technology regulation. In my view, it is gradually emerging as a new pillar of global governance, alongside trade, finance, climate and public health.
There is an important distinction, however.
Previous pillars of global governance were primarily designed to manage different forms of cross-border interdependence. Trade governance manages the movement of goods and services; international financial governance manages the movement of capital; climate governance addresses the accumulation of greenhouse-gas emissions; and global health governance seeks to prevent the international spread of infectious diseases.
AI is fundamentally different. What increasingly crosses borders is not simply goods, money, data or digital services. It is intelligence itself – the ability of machines to generate knowledge, make decisions, create content, accelerate scientific discovery, and increasingly perform cognitive tasks once reserved for humans.
That distinction has profound implications. Governing intelligence requires more than regulating products or markets. It requires institutions capable of addressing questions that no previous governance system has had to confront at this scale: How should increasingly autonomous systems be supervised? How can innovation be encouraged without allowing excessive concentration of technological power? How should the benefits of AI be shared between advanced and developing economies? How can legitimate national security concerns be balanced with scientific collaboration? And how should society ensure that AI strengthens, rather than weakens, public trust?
China’s recent emphasis on international cooperation adds an important perspective to this broader international discussion. Whether one agrees with every aspect of China’s approach is beside the point. The more significant contribution is its recognition that AI governance cannot be built by any single country acting alone. As AI systems become increasingly integrated into global supply chains, financial markets, scientific research and digital infrastructure, the cost of fragmented governance is likely to rise, because fragmentation generates incompatible standards, duplicated compliance requirements, and barriers to scientific collaboration, slowing innovation while weakening public trust.
This is not to suggest that countries will – or should – adopt identical regulatory systems. Different societies will continue to reflect different legal traditions, political institutions and development priorities. Diversity in governance is both inevitable and, in many respects, desirable. But diversity need not preclude cooperation. International financial regulation did not eliminate national banking systems. Climate agreements did not require identical domestic climate policies. Similarly, future AI governance will almost certainly combine national sovereignty with international coordination.
The larger lesson is that humanity has reached another institutional turning point. Every major technological revolution eventually requires new rules, new norms and, ultimately, new institutions. AI is unlikely to be an exception.
The real challenge before us is therefore not whether artificial intelligence will transform the world – it already has. The more important question is whether our institutions can evolve quickly enough to govern intelligence itself. If they can, AI may become one of humanity’s greatest engines of innovation, prosperity and sustainable development. If they cannot, the world’s most powerful technology may deepen the very inequalities, insecurities and geopolitical divisions that it also has the potential to overcome.
History will judge this generation not simply by how successfully it developed artificial intelligence, but by whether it had the wisdom to build the institutions needed to govern it. The countries that shape these institutions will influence not only the future of AI, but the future of international order itself.
Shouqing Zhu is a senior policy adviser at the United Nations Environment Programme Finance Initiative (UNEP FI). He has previously held senior policy and advisory roles at the World Economic Forum, UNICEF, UNDP and the World Resources Institute. The views expressed here are the author’s own.
