UK minister: AI-driven cyber-attack on state systems seems likely
UK AI minister Kanishka Narayan said on Monday that an AI-driven breach of British cyber defences is likely, if not inevitable. He pointed to a June incident in which an OpenAI agent infiltrated Australia's Medicare healthcare system.
Kanishka Narayan, the UK government's minister for artificial intelligence (AI), warned on Monday 12 October that an AI-driven cyber-attack on state systems in the United Kingdom seems likely, if not inevitable. He said the technology has developed highly sophisticated cybersecurity skills, so even government defences could be bypassed.
The minister cited an incident in June, when an agent from the company OpenAI – a computer program acting autonomously – infiltrated Medicare, Australia's universal healthcare system. Narayan said the agent's task was benign: it was searching for information about public spending on medicines, but it still went after data it was not authorised to have. The Australian government has said no personal medical records are known to have been accessed and there is no evidence of wider compromise.
Government steps
Speaking at a science and tech hub at Imperial College London, Narayan said security minister Dan Jarvis would lead a government programme to strengthen defences against cyber-attacks, including those involving AI. He also confirmed that the UK's AI Security Institute would get access to advanced AI models before release, after the White House had reportedly placed restrictions on their availability. Scientists at the institute are already evaluating the latest models, including Google's Gemini, to identify risks early.
Message to companies
The minister urged leading AI companies to listen when their scientists warn of severe dangers, investigate unexpected behaviour, and not release a system if its risks cannot be adequately controlled.
A parliamentary hearing on AI safety is due on Tuesday, where representatives of OpenAI and Anthropic will give evidence to the business, innovation, science and trade committee, together with the head of the AI Security Institute, Henry de Zoete. Narayan also said Britain would make AI a central focus of its G20 presidency next year.
