The Guardian reported that UK AI minister Kanishka Narayan will announce stronger cyber defences in a west London speech and call on frontier AI companies to increase their safety efforts. He will argue that Britain cannot rely entirely on companies to police their models and needs to build more of its own AI infrastructure. It will be his first speech since Andy Burnham made AI a cabinet-level brief.
Leaders of Finland, France, Germany, Australia, Singapore, the UAE and Ireland have signed a call for governments to coordinate common AI safety standards. The UK and US have not signed it. The prime minister told the UN in September that Britain would use its G20 presidency to champion a single set of global principles and standards.
Geoffrey Irving, a former chief scientist at the UK’s AI Security Institute, told the Guardian that AI companies lack security capable of preventing an adversarial country from stealing their models. He cited cybersecurity researchers who said they could theoretically access a great deal during a test conducted under OpenAI’s programme for ethical hackers. OpenAI says its protections against model theft include encryption and continuous monitoring.
