UK strengthens cyber defenses and calls for greater AI security measures

The UK government has announced plans to enhance its cybersecurity defenses and is urging leading AI developers to adopt more rigorous safety measures. AI Minister Kanishka Narayan is set to unveil these policies during a speech in West London. Narayan emphasized that the UK cannot rely solely on tech giants for self-regulation of AI models and stressed the need for domestic AI infrastructure. Narayan, appointed in July 2026, oversees the AI Security Institute (AISI), which has tested over 30 models with major firms like OpenAI and DeepMind. AISI’s 2025 report highlighted rapid AI advancements, with model success rates in cybersecurity rising to 50% and software engineering tasks reaching over 40%. The UK has also launched initiatives such as the Sovereign AI fund and the Cyber Growth Action Plan to bolster its AI and cybersecurity sectors. The government aims to regulate most AI systems during their use phase, reflecting growing concerns over AI safety and security.