Carney's AI for All Plan Meets a Divided United Nations
Canadian Prime Minister Mark Carney told the UN General Assembly on September 23, 2026 that the world faces a "once-or-twice-a-century rupture" and laid out Canada's 'AI for All' strategy.
Carney was in New York City from September 21 to 23 for the 81st Session of the UN General Assembly, according to the Prime Minister's Office announcement Prime Minister's Office. He spoke to the Assembly during the visit. The speech is listed in the official video record Prime Minister's Office.
On September 23, he outlined 'AI for All' in New York CanadianPM. Reporting from the Assembly hall described the rupture language as a broad argument about world order, not just a technology pitch The New York Times.
That split is about speed versus shared rules. Should AI move fast, or should many countries agree on limits first? Multilateral rules here means limits set together by many countries. Think of it like traffic rules for a new highway. On September 22, UN Secretary-General Antonio Guterres used his final Assembly speech to call for curbs on AI Reuters. That followed his September 16 warning that fast-moving AI creates risks Reuters.
The United States took a different view. On September 13, Donald Trump called AI critics "very negative forces" raising scenarios that will not happen Reuters. On September 14, Carney and Trump were reported to share doubts about calls to slow AI Toronto Star. That reporting put Ottawa closer to Washington than to Guterres on pace, even as Canada stayed active in UN work.
Canada also works country-to-country on AI. On February 14, 2026, Canada and Germany signed a Declaration of Intent on AI on the sidelines of the Munich Security Conference. Ottawa said then it would welcome Germany as Country of the Year at the All In conference in Montreal in September 2026 Innovation, Science and Economic Development Canada.
The broader context here is that AI rules are splitting apart. Guterres used his last Assembly platform to push for common limits. Trump rejected caution in strong terms. Carney presented Canada as focused on spreading access, using the label 'AI for All,' while questioning calls to slow down. For diplomats, that three-way difference matters more than any one speech.
Looking at what this means for talks between countries, Canada is trying to hold two ideas at once. The first is national strength, built with partners like Germany through joint statements and tech partnerships. The second is wide sharing, spoken in the Assembly language of inclusion. That pairing lets Ottawa talk to Washington about moving fast and to other capitals about access, without backing binding limits.
In my view, the test is what gets built next. A declaration of intent sends a political signal. An Assembly speech brings attention. On their own, neither creates enforcement, standards, or purchasing rules. Close watchers will look for funding plans, deals to share computing power, joint safety tests, or shared positions in standards bodies. They will also watch whether Guterres' successor keeps pushing limits after his final speech, and whether the reported Canada-U.S. agreement on speed leads to joint regulation or stays at the level of words.
Looking ahead, the stakes for the UN are direct. If big technology powers treat AI mainly as an economic race, the General Assembly becomes a stage for competing plans rather than a place that sets rules. Carney's rupture line recognizes that change. It does not settle it.
