Today, I update the House on a significant milestone in the UK's AI sector. The country has seen a 688% increase in AI companies over less than a decade, with scale-ups raising almost double that of France, Germany, and other European countries combined. However, the rapid advancement in AI presents both immense opportunities and potential risks. Last week, we convened an international summit at Bletchley Park, where leaders from 28 countries and the EU signed the 'Bletchley declaration', agreeing on principles for human-centric, trustworthy, and responsible use of AI. This includes measures to protect human rights, ensure transparency, fairness, accountability, safety, oversight, ethics, bias mitigation, privacy, and data protection. Additionally, an international agreement was secured on pre-deployment testing of models by governments, led by the Turing prize winner Yoshua Bengio for a state-of-the-art report. The UK will establish a permanent AI safety institute to conduct third-party testing, with funding confirmed for the rest of the decade. Future summits are scheduled in South Korea and France, ensuring continuous international collaboration. Beyond this summit, our White Paper ensures agility and responsiveness in legislation as risks evolve, without rushing into inadequate regulation. Pre-deployment testing is a critical achievement, essential to mitigate AI risks responsibly. Companies also committed to publishing safety policies, fostering transparency and competition. The minister emphasises the transformative potential of AI while stressing the importance of safety governance to prevent misuse and uphold societal trust.