UK pledges global AI safety standards for 2027 G20 presidency
British Prime Minister Andy Burnham announced that the United Kingdom will place artificial intelligence safety at the centre of its G20 presidency in 2027. Speaking at the United Nations General Assembly, Burnham said the goal is to agree on a single set of global principles and standards for AI development, focusing on transparency, access and preparedness.
A defence partnership with the United States
The British leader also announced a new bilateral agreement on AI security. "We are developing a new AI-based defence partnership with the United States to protect our critical national infrastructure," Burnham told the Assembly. The partnership aims to defend essential systems — power grids, telecommunications, financial networks — from cyber attacks enhanced by AI capabilities.
The announcement comes at a time when several G20 members are strengthening their own regulatory frameworks. The European Union's AI Act became fully applicable in August 2026, while the United States has pursued a voluntary approach under the current administration. China maintains strict pre-launch security assessments for generative AI services.
Governments, not tech companies, must decide
Burnham argued that fundamental decisions on AI should rest with elected governments rather than technology companies. He drew a parallel to the coordinated G20 response to the 2008 financial crisis, when common principles were established and then applied nationally. The British Prime Minister said the same approach is needed for AI governance.
The UK has positioned itself as a bridge between the EU's regulatory framework and the US innovation-driven model. Britain established an AI Security Institute (AISI) to examine advanced AI models for capabilities and potential misuse. However, questions have been raised about voluntary compliance after reports that at least one company declined to submit a model for evaluation.
The Italian context
For residents in Italy, the British initiative adds another layer to an already complex landscape. In August 2026, the Italian Council of Ministers approved legislative decrees to align national law with the EU AI Act, including provisions on cybersecurity and data centres. Italy has also deepened cooperation with the United States through the "Pax Silica" agreement on critical minerals signed in September 2026, and extended a cybersecurity memorandum with the US Department of Homeland Security until January 2027.
Tensions ahead
Diplomatic sources note potential friction in the harmonisation of standards. The United States has rejected what it terms "globalist schemes" to control AI, preferring bilateral arrangements. Burnham's vision of a unified global framework may therefore face resistance from Washington as much as from Beijing. The operational details of the new UK-US partnership and the precise G20 calendar have not yet been disclosed.
What happens next
The British G20 presidency begins in December 2026. Until then, Italy and other EU members will continue implementing the AI Act's requirements, which include mandatory transparency for AI-generated content and the establishment of regulatory sandboxes by August 2026. The UN Secretary General has called for a Global AI Fund and stronger safeguards for children, while an international scientific panel has warned that current AI defences are "unravelling".
