There is a single thread running beneath the most consequential technology stories of the moment: the regulation of artificial intelligence in Europe and the Americas, digital money, data sovereignty, the chip war, platform work, elections threatened by synthetic content. If one question sits under all of them, it is this: can you govern a technology that respects no borders, when the parties who would have to agree on the rules will not? Global AI governance is the arena where that question becomes explicit — it does not resolve the puzzle, but it states it with unusual clarity.
The starting point is an admission from the very top. “AI is moving at the speed of light. No country can see the full picture alone,” UN Secretary-General António Guterres has said, stressing the need for shared understanding to build effective guardrails. His diagnosis is also the diagnosis of this larger story: the technology advances faster than the rules, and the rules, moreover, fail to become common. Guterres has put it in stark terms — paraphrasing his framing, the choice is between governing AI and being governed by it — and he has pushed mechanisms to attempt it: a Global Dialogue on AI Governance and an Independent International Scientific Panel on AI, the latter conceived, in the UN’s own analogy, to be to artificial intelligence what the Intergovernmental Panel on Climate Change is to global warming.
A summit that mapped the divide
The real state of that aspiration was exposed in February 2026, at the AI Impact Summit in New Delhi — the fourth in the annual series, and notable for convening an extraordinary array of heads of state, corporate leaders, and technologists in a major emerging economy. More than 80 countries backed a vision of “safe, secure and trustworthy” AI. But the fact that defined the summit was a refusal: the United States, home to most of the world’s leading AI companies, declared that it “totally” rejected global governance of artificial intelligence.
The American position, voiced by White House technology adviser Michael Kratsios, is coherent and worth stating precisely: that AI “cannot lead to a brighter future if it is subject to bureaucracies and centralised control,” that excessive international regulation could — in the words used at the previous Paris summit — “kill a transformative sector,” and that the climate of fear around the debate should give way to one of opportunity. This is not a careless stance; it reflects a genuine conviction that innovation flourishes best without multilateral constraints. But its effect is unambiguous: without the country where the bulk of AI development is concentrated, any global governance is born limping. It is the same logic visible in the chip war and in data sovereignty — the technology is planetary, but the power to regulate it is concentrated in a few hands that do not necessarily wish to be regulated.
Two visions contend for the future
The underlying debate has two poles, and both hold part of the truth; the contest between them will define the coming years.
The global-governance view holds that a technology with planetary effects — on employment, information, security, rights — requires common rules, much as the climate or civil aviation do. Its advocates argue that leaving AI to each country, or worse, to each company, guarantees a race to the bottom in which safety and ethics are sacrificed to speed and competitive advantage. Science-based governance, they contend, is not a brake on progress but a way to make it safer, fairer, and more widely shared. The UN initiative itself was approved by consensus in the General Assembly after months of negotiation led, notably, by Spain and Costa Rica.
The free-market view responds that innovation is fragile and that overregulation — especially the international and bureaucratic kind — can smother it; that global frameworks move at a pace incompatible with the technology; and that it is preferable to let each country experiment with its own rules, allowing competition and practice to reveal what works. Even some who reject binding global structures, the US among them, describe both regulatory and non-regulatory frameworks as a “necessity” and signal openness to voluntary, nonbinding cooperation. France’s president, in New Delhi, tried to occupy the middle ground: Europe, he argued, “is not blindly focused on regulation” but is “a space for innovation and investment” — a safe one.
It is not for this outlet to decree which vision is right; the outcome will depend on a political and geopolitical struggle that is only beginning. What can be stated is that the tension between the two — common rules versus decentralised experimentation — is exactly the one that recurs across nearly every adjacent story, from AI regulation to the fragmented map of cryptocurrency rules.
Where the developing world stands
For most of the world’s countries — those that neither build the frontier models nor own the infrastructure they run on — the decisive question is whether they have any voice in this conversation at all. The picture is mixed. There are signs of participation: the UN’s Global Dialogue was shepherded through months of negotiation by Spain and Costa Rica, the latter a small country positioning itself, as in other arenas, to play in the league of digital governance. Brazil signed an agreement with India on critical minerals and rare earths essential to the tech industry, staking a place in the very supply chains contested in the chip war. And several emerging economies are climbing fast in measures of AI maturity.
But optimism should be tempered. These countries arrive at the table with unequal weight: they do not produce the frontier models, they depend on others’ infrastructure, and their voice, though present, lacks the heft of the major powers. A Costa Rica co-sponsoring a dialogue, or a Brazil signing mineral deals, is meaningful but modest against the scale of decisions taken in Washington, Beijing, or Brussels. There is a real opportunity here to be more than a rule-taker — but seizing it demands coordinated, strategic action, which does not always materialise. Analysts also warn of a quieter risk: that parallel initiatives, like the New Delhi Declaration and the UN’s Geneva-bound dialogue, evolve in isolation rather than in concert, deepening the very fragmentation they were meant to cure.
What this story reveals about all the others
Global AI governance works as a synthesis. Every adjacent theme — the AI that adjudicates credit or curates content, the digital money that challenges monetary sovereignty, the data that lives under foreign law, the chips two powers fight over, the platforms that direct labour, the elections menaced by deepfakes — is, at bottom, a variation on the same tension: a powerful, borderless technology confronting national institutions that are slow and fragmented in their attempts to govern it. Global governance is where that tension appears in its purest form, because there the object to be governed is the technology itself, and the would-be governing subject — the international community — cannot constitute itself as one.
The verifiable fact is that the world recognises the need to govern artificial intelligence but cannot agree on how; that the leading technological power rejects global governance while the UN and dozens of countries promote it; and that the developing world is beginning to seek a voice it does not yet have the weight to impose. Whether humanity will govern this technology fairly and in common, or leave it to be defined by the interests of those who control it, depends on decisions only now being taken: on whether the major powers find common ground despite their rivalry, on whether multilateral bodies achieve real relevance, and on whether today’s peripheral regions organise to be heard. As in every story of this kind, what is decisive is not the technology itself — which will advance, with rules or without them — but whether human societies manage, this time, not to arrive too late.