Part Four: Who Gets to Write the Rules of the AI Age?

Perhaps the most important question emerging from this year's United Nations debate over artificial intelligence is not whether AI needs rules. A remarkable number of governments now appear to accept that some form of international governance will be necessary.

The deeper argument concerns who gets to create it.

The United Nations itself offers a warning from history. Much of today's international architecture emerged after the Second World War, when a comparatively small number of victorious powers possessed enormous influence over the institutions being created. The clearest example is the Security Council, where the United States, China, Russia, Britain and France still possess permanent seats and veto powers inherited from the political settlement of 1945.

Africa still has no permanent seat.

African governments raised that issue forcefully again during this year's General Assembly. The Democratic Republic of Congo reiterated the African Union's demand for two permanent African seats, with the same privileges as existing permanent members for as long as those privileges continue to exist. Supporters of reform argue that a continent representing more than a quarter of UN member states cannot indefinitely remain outside permanent membership of the institution responsible for international peace and security.

There is an important connection between that debate and artificial intelligence.

Many developing countries appear determined not to allow the international governance system for one of the twenty-first century's most transformative technologies to be created in the same way that major twentieth-century institutions were. They do not want rules written primarily in Washington, Beijing, Brussels or Silicon Valley and subsequently presented to the rest of the world as a completed framework.

India wants developing states directly represented in AI governance. South Africa wants African countries participating both in rulemaking and in the economic value generated by AI. The DRC wants greater international investment in capacity so that poorer countries do not become permanently dependent upon a small collection of technological powers. China increasingly presents itself as an advocate of wider access to AI technologies. The United States continues to view technological leadership partly through the lens of strategic competition. Singapore is beginning to ask whether the eventual answer could require a treaty-like international framework.

These positions are not identical, and sometimes they directly conflict. Yet they reveal the shape of an emerging diplomatic argument. Artificial intelligence is rapidly becoming another arena in which questions of development, security and international representation intersect.

The practical problems are formidable. Should governments have rights to inspect the world's most powerful AI systems? At what point does a commercial model become a matter of national security? How should autonomous weapons be regulated? Should military commanders always be required to approve lethal decisions? Can countries realistically restrict the spread of powerful models once they are developed? Who is legally responsible when an autonomous system makes a disastrous decision?

Then there is the economic question. What prevents the world from dividing into a small group of states possessing advanced computing infrastructure and a much larger group dependent upon them? Access to artificial intelligence may eventually become as strategically important as access to energy, capital or telecommunications.

The United Nations is not perfectly designed to answer these questions. Its fundamental institutional architecture is more than eighty years old, and many of its mechanisms struggle even with the problems they were originally designed to solve. Yet the UN has one enormous advantage when confronting an issue such as artificial intelligence: almost every recognized state is represented there.

That matters.

The future governance of AI cannot realistically be determined only by the companies building the models or by the governments of the countries in which those companies happen to be located. The technology will affect economies, education, warfare, information, scientific research and public administration across virtually every country.

The 128 delegations that discussed AI during this year's General Assembly were not agreeing about what the rules should be. In fact, the disagreements are becoming increasingly obvious.

But they were signaling something important: they expect to have a voice.

Future historians may therefore look back at General Assembly week in 2026 for reasons that were not obvious from the television coverage. The memorable images may have involved diplomatic confrontations and walkouts. Yet underneath those familiar scenes, an entirely new field of international diplomacy was beginning to take shape.

The struggle over artificial intelligence is no longer simply a competition over who can build the most capable machine.

It is becoming a much larger argument over who gets to shape the international order in which those machines will operate.