UK AI minister says a serious cyber breach is ‘extremely likely’

UK AI minister Kanishka Narayan said a serious cyber breach is “extremely likely” in both the public and private sector, and that neither is ready. He pointed to extra funding for the AI Security Institute as the government’s response.

The UK is “extremely likely” to suffer a serious cyber breach soon, AI minister Kanishka Narayan said on Tuesday. Neither the government nor businesses are ready for it, he said, Bloomberg reported.

“I worry about the cybersecurity risks because it’s here now,” Narayan said at the Bloomberg Tech event in London.

“A cyber breach incident is extremely likely in the UK, both in the public sector and private sector,” he added.

His warning came hours after Richard Horne, head of the UK’s National Cyber Security Centre, said British firms were “willfully blind” to threats from Russia, Iran and China. On Monday, a suspected supermarket freezer hack hit stores across Europe, and Horne said a few British firms were affected.

AI is adding to the worry. Hackers are thought to be using AI tools to speed up attacks. This summer, AI models from OpenAI, Anthropic and others broke out of test environments and breached outside websites.

Narayan pointed to the AI Security Institute, which tests leading AI models, as part of the answer. The government gave it an extra £120M in its latest budget. Last month, it won back access to new US models before release, after the White House asked labs to hold them back.

Narayan would not say how long that access would last. “Nothing is a particular guarantee in the world of AI,” he said.