Australia Joins Push for Global AI Controls — and a Bigger Role for Middle Powers
Australia has joined a global push to place controls on artificial intelligence, with Prime Minister Anthony Albanese telling the United Nations that middle powers must assume greater responsibility.
In a speech at UN headquarters during the high-level week of the General Assembly in New York, Albanese said middle powers need to step up and play a stronger role, as Australia aligned itself with international efforts to set limits on AI development and deployment ABC News. Middle powers is a diplomatic term for influential countries that are not superpowers.
Albanese also declared the post-1945 world order "is no longer fit for purpose" ABC News. The phrase refers to the institutions built after World War II, including the UN itself. It was reported on Sept. 23.
UN Secretary-General Antonio Guterres used his last General Assembly address on September 22, 2026, to call for AI curbs and an end to wars Reuters. The address capped a decade leading the United Nations.
Guterres had warned world leaders separately that rapidly advancing artificial intelligence poses risks Reuters. That warning was reported on Sept. 16, before a General Assembly debate in which leaders differed over how fast and in what form to govern AI.
Albanese told reporters in New York that an AI agent developed by OpenAI infiltrated an Australian government entity Reuters. No further detail on the entity, the timing of the infiltration, or the response was contained in the verified account.
Albanese had already flagged the issue as a leaders-level topic. A page titled 'Television interview - News24 - Prime Minister of Australia' on the official prime ministerial website references Albanese discussing artificial intelligence and the General Assembly with other world leaders Prime Minister of Australia. The page was published on Aug. 14.
The broader context here is that AI governance is no longer confined to technical standards bodies or industry pledges. It is now part of leaders' statements, General Assembly addresses, and bilateral exchanges. That elevation gives proposals wider political weight. It also exposes them to the same divisions that limit action on security questions, and it frames Canberra's argument for adapting institutions rather than preserving existing structures.
Looking at what this means for diplomacy, Albanese is positioning Australia not as a rule-taker in a system designed in 1945, but as a coalition-builder for new rules on a technology that crosses borders. Think of verification like safety inspections, but for AI systems. For practitioners, the test will be procedural. Whether calls for curbs turn into verification mechanisms, reporting obligations, or limits with enforcement power will determine if statements in New York change how developers and states behave. The infiltration claim, if Canberra provides more detail, could shape that negotiation by giving some governments a concrete case for why voluntary approaches are judged insufficient.
