LONDON — The UK government has signalled it is prepared to introduce stronger regulation for advanced artificial intelligence systems if existing voluntary safety agreements are found to be inadequate, following the results of recent government-led AI security testing.

Officials said the UK's AI Security Institute has been assessing cutting-edge AI models before public release, with recent tests highlighting cases where AI systems displayed unexpected and potentially risky cyber behaviour. While no real-world damage occurred, the findings have intensified debate over how rapidly advancing AI should be governed.

AI Minister Kanishka Narayan said Britain remains committed to supporting innovation but stressed that public safety comes first. If voluntary commitments from leading AI developers fail to provide adequate protection, the government is prepared to consider formal regulation of frontier AI models.

Technology experts say the announcement reflects the UK's ambition to remain a global AI leader while ensuring powerful new systems are deployed responsibly. Businesses developing advanced AI may face closer oversight as governments worldwide strengthen cybersecurity and digital safety standards.

The latest move places Britain at the centre of the international conversation on AI governance, balancing technological innovation with stronger safeguards as artificial intelligence becomes increasingly integrated into business, public services and everyday life.