UK outlines plans to use frontier AI for cyber defence

The government said it would extend a programme using frontier AI to strengthen cyber defences, with the aim of protecting public services in the short term and supporting wider public-sector adoption over time.[1] It reported that two pilots had helped identify and fix more than 1,000 potential vulnerabilities across government systems and detect potential cyber threats.[1] The pilots helped identify and fix more than 1,000 potential vulnerabilities across government systems and detect potential cyber threats.[1]

Separately, the Government Cyber Coordination Centre led weekly, in-person hackathons that used frontier AI to scan government public-code repositories.[4] The stated aim was to find and mitigate previously unidentified vulnerabilities before they could be exploited.[4] The centre is a partnership between the National Cyber Security Centre and the Department for Science, Innovation and Technology.[4] The government said two pilots helped identify and fix more than 1,000 potential vulnerabilities, and separately described the hackathons’ scanning work.[1][4]

The government said it would consider stronger AI-threat protections through the Cyber Security and Resilience Bill regime, including ways to identify and report serious incidents and near misses.[1] New work is also underway to prepare Britain for serious incidents involving autonomous AI agents.[1] Narayan argued that responding to AI risks takes more than passing laws and requires institutions with deep expertise; he also said AI machines should leave human beings in charge.[3] He warned that failures or malicious use could have more serious consequences as frontier systems become more capable and autonomous.[1]

Narayan called for a whole-society approach to resilience and said Britain could not rely entirely on companies to police their models.[1][2] Leaders from Finland, France, Germany, Australia, Singapore, the UAE and Ireland signed a call for governments to coordinate common AI safety standards; the UK and US did not sign it.[2] The prime minister told the UN in September that Britain would use its G20 presidency to champion common global AI principles and standards.[2]