Multilateralism has fallen on hard times lately, especially at the United Nations. The UN Security Council couldn’t stop Russia’s invasion of Ukraine, ceasefires in Gaza and Lebanon remain elusive, and subsequent COP summits have failed to spur enough concrete action to meet global climate targets. Not only are the UN’s own Sustainable Development Goals off-track; in many cases, progress toward meeting them has reversed. The UN’s foundational commitments to peace, security, and cooperation feel very foreign at a time when multiple wars are raging, protectionism is on the rise, and the world is splitting into rival blocs.
But despite this geopolitical recession, global cooperation is still possible. The UN General Assembly’s first Summit of the Future on September 22-23 tested the organization’s ability to tackle one of the world’s biggest transnational challenges: artificial intelligence. Surprising as it may be, the UN passed.
It is no exaggeration to say that AI has spurred one of the fastest and most robust policy responses in living memory. Barely a year ago, UN Secretary-General António Guterres invited representatives from government, the private sector, and civil society to recommend how the world might govern AI in the service of humanity. He knew that the world’s ambition to govern AI could fall flat, much like the initial response to climate change. The existing approaches were already too fragmented, and most left out the Global South, with 118 countries party to no AI governance framework at all.
Together, we served as rapporteurs for the secretary-general’s High-Level Advisory Body on AI, which was established to meet this worthy challenge. Reflecting the world’s diversity, its 39 members came from every continent, and included representatives from government, academia, civil society, and the technology industry.
This was the first genuinely global effort to govern AI, and we are pleased that several of our recommendations were taken up in the Global Digital Compact, a comprehensive governance framework that UN member states adopted last month. Reaching this new agreement required overcoming all the very real differences that separate the United States, China, Europe, and the Global South, as well as governments and the private sector (especially technology companies).
For example, one of our recommendations – which has been approved in principle for implementation – is to establish an International Scientific Panel on AI. We started from the premise that to govern an issue as complex as AI, we should have a common understanding of the technology and its potential risks and effects across countries and cultures.
We learned this lesson the hard way from climate change. While many now debate how to address the climate crisis, there is no serious debate over whether we should address it; the evidence provided by the Intergovernmental Panel on Climate Change is overwhelming. A similar intergovernmental panel on AI would undertake the difficult but fundamental work of analyzing the ongoing developments in AI technology, thereby giving policymakers a factual, independent foundation to inform debates, goals, and policy decisions.
But what we are most enthusiastic about is the prospect of ensuring that AI benefits everyone. Unlike climate change – where there are zero-sum politics and serious short-term trade-offs between lowering emissions, fostering economic growth, and achieving equity (with powerful vested interests that oppose a post-carbon transition) – AI is the rare transnational issue with positive-sum solutions. If shared safely and made to respect international law and fundamental freedoms, AI should not pose an existential threat to incumbent governments and companies. Instead, it should catalyze win-win opportunities.
There is tremendous demand for technologies like AI, as well as excitement over its potential to help us meet all sorts of objectives, including those enshrined in the SDGs. From public health and education to economic growth and climate mitigation, AI can be a game-changing technology. But without the infrastructure and mechanisms to oversee its transformative growth, it could drive further global divergence, with the poorest and most vulnerable populations once again being left behind. We are determined to prevent that.
That is why, in addition to forming a common knowledge base, we have recommended initiatives that enhance all countries’ and communities’ access to AI. From talent and standards to data and funding, the UN and its partners can help address gaps in resources and infrastructure to ensure that no one is left behind from the AI revolution.
Of course, there are some who question the UN’s role in governing AI, and governance must take place at the nation-state level as well. The companies developing AI models also are creating standards. But like the internet before it, AI’s potential makes it a global public good (as is AI safety). The UN is the only truly global body with the legitimacy to convene the world’s governments and AI stakeholders, and the ability to guarantee any resulting agreements. That starts with getting the world on the same page – not to compel governance, but to align around the nature and scale of the opportunity and challenges. With the right vision, tools, and political leadership, we can deploy the resources to ensure that AI lives up to its promise.
From climate change and public health to nuclear proliferation, the world has turned to the UN to solve its most complex problems. Armed conflict, humanitarian disasters, environmental crises, and economic woes highlight the international community’s frequent failure to rise to the challenges the world faces. But as we grapple with our most revolutionary and potentially disruptive technology yet, the Global Digital Compactproves that there is still hope for multilateralism in a geopolitically fragmented world.
Project Syndicate
