UK Government Turns Frontier AI on Its Own Systems to Find Flaws Before Attackers Do

Speaking at Imperial College London on October 12, UK AI minister Kanishka Narayan outlined a set of initiatives to protect the country from emerging AI threats. A central plan is the urgent use of frontier AI to find weaknesses in government systems before attackers exploit them. Narayan said the government's central vulnerability-monitoring service has already cut the time taken to fix identified weaknesses from 50 days to eight.

The approach builds on internal hackathons run by the Government Cyber Coordination Centre, where frontier AI helped find more than 400 software flaws in public sector IT. Those flaws were subsequently patched. Narayan also said the Department for Digital, Culture, Media and Sport, the National Cyber Security Centre and the AI Security Institute will work together to develop robust ways to respond to and contain incidents involving agentic AI. This follows cases in which frontier models have hacked organisations, including governments, during testing.

Security minister Dan Jarvis will lead a cross-government program to strengthen defences against cyber-attacks, including those involving AI, and a National Cyber Action Plan will be published soon. Amendments to the Cyber Security and Resilience Bill are also planned, covering identifying and reporting serious incidents and near misses, reducing AI-driven risk in essential services, and ensuring serious AI incidents can be properly reviewed. Many details are still being worked out.