Part Three: How Artificial Intelligence Reached the Top of the Global Agenda
Among all the numbers generated during this year's United Nations General Assembly, one deserves particular attention. Four years ago, only five national delegations mentioned artificial intelligence during their General Assembly speeches. This year, 128 did. According to the President of the General Assembly, AI was discussed by more delegations than either climate change or human rights.
That does not mean governments suddenly consider climate or human rights unimportant. It means that artificial intelligence has crossed an important political threshold. It is no longer being treated primarily as a technology-sector issue. Governments increasingly understand AI as economic policy, military policy, development policy, foreign policy and national security policy at the same time.
The United States presented one of the clearest expressions of AI as strategic competition. President Donald Trump argued that leadership in increasingly powerful forms of artificial intelligence would provide enormous national advantages and emphasized the importance of maintaining American leadership over China. The underlying message was familiar from earlier technological rivalries: advanced AI is becoming a source of national power.
China framed the question differently. Vice President Han Zheng emphasized the need to ensure that artificial intelligence did not become a capability reserved for wealthy countries. China increasingly presents AI development through the language of access, economic development and narrowing the technological divide between the developed and developing worlds. This is also part of a wider geopolitical competition. The struggle over AI may eventually be less about which country owns the single most powerful model than about which technological ecosystem much of the world chooses to join.
India concentrated on another question: who writes the rules? New Delhi argued for an inclusive international framework in which member states remain central and developing countries participate directly in AI governance. India has also emphasized locally relevant applications and wider access to the computing and technical resources needed to develop AI rather than simply consume products built elsewhere.
African governments made the inequality issue especially explicit. The Democratic Republic of Congo warned that the existing digital divide could become an "intelligence divide." It is a useful phrase because the next great technological inequality may no longer be measured simply by whether a country has access to the internet. It may depend on whether it possesses affordable electricity, data centers, computing power, advanced semiconductors, skilled engineers, high-quality datasets, capital and access to powerful models.
South Africa similarly argued that African countries must participate in creating the rules governing artificial intelligence rather than merely providing resources to the countries and corporations developing it. This connects AI governance with a much older development question. Africa possesses substantial reserves of the critical minerals required by modern technology and a rapidly growing young population, yet many African governments are determined not to repeat a familiar economic pattern in which raw materials leave the continent while most of the value is created elsewhere.
Singapore offered one of the week's most ambitious ideas by suggesting that the international community begin considering a United Nations Framework Convention on AI Safeguards. The phrase is important because framework conventions are used when governments recognize a problem too large to solve through a single agreement. Climate change provides the obvious precedent: states first establish principles and institutions and then negotiate more specific rules over time.
Singapore also moved the discussion directly into the territory of international security by asking a simple question: would we want an artificial intelligence controlling the nuclear button? The Holy See approached the same issue from an ethical perspective, emphasizing that decisions involving lethal force must remain subject to meaningful human control and accountability.
European governments have tended to emphasize transparency, regulation and accountability. Britain, meanwhile, has increasingly connected AI with foreign policy, economic strategy and defense. Once governments begin talking about artificial intelligence in relation to autonomous weapons, information warfare, biological threats, critical infrastructure and great-power competition, the subject has plainly left the technology conference behind.
We are therefore beginning to see the outline of a new international policy field. For much of the past decade, debates about AI concentrated on innovation, automation, jobs and ethics. Those debates remain important, but governments are now confronting much larger questions. Who controls the most powerful models? How should military AI be governed? Can dangerous capabilities be inspected? How do states prevent advanced systems reaching terrorist organizations or hostile actors? Who is responsible when autonomous machines make harmful decisions?
These questions increasingly resemble the questions international institutions confronted during the nuclear age.
That does not mean artificial intelligence is equivalent to nuclear weapons. It means the governance problem is becoming similar. Governments recognize enormous potential benefits while simultaneously fearing what may happen if competitors acquire capabilities they do not possess.
Once that happens, regulation becomes inseparable from geopolitics.
Artificial intelligence has reached the top of the United Nations agenda because it is increasingly becoming part of the architecture of international power.
