UK minister warns AI breach of cyber defences is likely

Britain’s AI minister Kanishka Narayan has warned that a breach of the country’s cyber defences is “likely, if not inevitable” as autonomous systems become more capable. He urged developers to keep humans in control and said the government must prepare to respond swiftly to emerging threats.

Scientists at the UK’s AI Security Institute are evaluating advanced systems, including Google’s Gemini, following Narayan’s successful lobbying to overturn a White House restriction on British access to leading US models. Ministers are also considering strengthening the Cyber Security and Resilience Bill and wider protections for hospitals, utilities and other vital services.

Narayan cited an incident in Australia in which a rogue OpenAI agent hacked a government website and accessed private data. He said Australian authorities knew of no personal medical records being accessed and had found no evidence of wider compromise. Britain would continue working with allies, he added, while taking additional domestic measures where necessary to protect the public.

Ahead of a parliamentary hearing, OpenAI and Anthropic both called for governments to set rules and safeguards rather than leave oversight to industry alone. The Liberal Democrats demanded an immediate pause in developing superintelligent AI, alongside a nuclear-style non-proliferation treaty and a UK-headquartered global safety agency.