Chapter 1
The United Nations came to artificial intelligence by resolution
The United Nations came to artificial intelligence by resolution, and its resolutions bind no one. In March 2024 the General Assembly adopted its first on AI without a vote: a US-led text promoting 'safe, secure and trustworthy' AI, co-sponsored by more than 120 states and recommendatory only. In August 2025 it went further, creating two bodies meant to interlock: a 40-member Independent International Scientific Panel on AI to assess the evidence, on the model of the climate panel, and an annual Global Dialogue to decide what to do with it. The Panel's first report came in July 2026.
Chapter 2
Meanwhile the clubs of the richest economies brought the AI companies…
Meanwhile the clubs of the richest economies brought the AI companies into the room. At the G7 summit in Évian on 17 June 2026, the chief executives of OpenAI, Anthropic and Google DeepMind joined the leaders for a working session, only the second time in the group's fifty-two years that leaders met private-sector representatives formally during a summit; the outcomes document published no commitment from any of them. At the G20's innovation ministerial in Chapel Hill in September, the American hosts gave the on-stage interviews to technology chiefs, who used them to ask ministers for lighter regulation.
Chapter 3
September at the UN showed how far apart the governments are
September at the UN showed how far apart the governments are. Before the General Assembly's high-level week, Secretary-General António Guterres said voluntary efforts to slow AI 'will not be sufficient if they are isolated, unverifiable or unevenly applied'. Finland's president and Norway's prime minister launched a Call for Control of Frontier AI Models, signed at launch by leaders of twenty countries and the European Commission and within a week by nine more, France among them, asking UN members to explore an international institution able to set standards and verify frontier models; it sets no deadline and creates nothing, and the United States, China, Britain, Japan and India have not endorsed it. At a Security Council meeting France convened, the chief executives of OpenAI and Anthropic asked for common rules, and the White House science adviser said the risks were no reason for new global governance structures; the day before, President Trump told the General Assembly the United States 'rejects any attempt to construct a globalist scheme to control' AI. No text was adopted. Canada's AI minister said he had raised a G7 'technology stability board' to evaluate models before release; no other government has confirmed it, and nothing has been launched.
Chapter 4
In October the finance track took AI up on the same terms
In October the finance track took AI up on the same terms. At the IMF and World Bank meetings in Bangkok, G7 finance ministers and central bankers agreed to rehearse AI-driven cyberattacks on the financial system, the Financial Stability Board issued twelve practices for financial firms that it says are not a binding standard, and the US Treasury, as G20 president, pressed to have the IMF's steering committee call the technology 'super intelligence'. What the multilateral system has built so far is a panel, a dialogue, rehearsals and guidance; the verification body the Call for Control asks for and Canada's proposed board exist only as proposals, and the two largest AI powers are outside the one call that asks for verification. The calendar ahead runs through the G20 leaders' summit in Miami in December, the Global Dialogue in New York in May 2027, an AI summit in Geneva in June 2027, and a British-hosted G20 in Manchester that November.
And now
The multilateral system keeps adding rooms for talking about AI, and none of them can bind anyone. On 23 October ten Security Council members, Britain and France among them, set up an informal expert group on technology, peace and security, co-chaired by Greece and Latvia, to brief the Council before mandates are renewed. The United States, China and Russia are not founders. It is not a Council decision and has no mandate, budget or staff.
On 28 October Finland’s president and Norway’s prime minister met leaders of 31 countries online about their Call for Control of Frontier AI Models. Nothing was adopted; officials will explore “potential international mechanisms”, with no dates. The Call, which asks states to explore a body able to verify frontier models, has 29 countries and the European Commission behind it, France among them. The United States, China, Britain, Japan and India have not endorsed it.
The UN named the co-chairs of its 2027 dialogue on AI, and its disarmament institute announced a forum on military AI built on voluntary guidelines. The finance track produced a chair’s statement and practices that bind no one.
Where it stands today →