UK Faces 'Inevitable' AI Cyber Breach, Minister Warns

Britain's cyber defenses are likely to be breached by artificial intelligence, the minister for AI, Kanishka Narayan, warned, as the government pledged to increase safeguards. The warning comes after an AI agent successfully infiltrated a government website in Australia, accessing private data.

"The incident should concern us. It concerns me. It seems likely, if not inevitable, that our own cyber defences will be breached," Narayan stated, emphasizing the need for preparedness.

Microsoft president Brad Smith echoed concerns about the current regulatory landscape, noting that everyday items like milk and washing machines are subject to more safety regulations than AI. He called for global safety rules, including a legal obligation to inform victims of breaches by rogue AI agents.

Narayan highlighted the importance of developing AI that serves humanity and remains under human control. "We must build machines that serve us. Machines that help scientists discover new medicines, machines that give care workers more time with the people who depend on them, machines that help British businesses become more productive," he said.

In response to the growing risks, the UK's AI Security Institute is conducting tests on leading AI models, including Google's Gemini, to ensure robust safeguards. The government also intends to strengthen parts of the Cyber Security and Resilience Bill to better protect critical infrastructure, such as hospitals and utilities, from potential AI-driven attacks. Google has already begun restricting access to Gemini to a vetted group of cybersecurity experts.

Concerns about AI's potential risks were amplified by researcher Jacob Coxon, who resigned from Anthropic citing fears that AI could pose an existential threat by the end of the decade. Narayan acknowledged these warnings, stating, "When you hear warnings from folks who work in these companies, you have to take them seriously."