The UK budgets for agent security as cyber scenarios grow more demanding

It also provides an additional £120 million for the AI Security Institute across 2026–27 to 2029–30. These are funding commitments for distinct capabilities, not evidence that the money has already been spent or that either organization can prevent every advanced attack. [1] A November 6 ministerial answer confirms that the capability was first announced in September and that operational details are still forthcoming. [3]

NCSC’s November 3 report sets out three plausible futures for AI-assisted cyber intrusion. Its judgments include a growing risk of AI-powered malware and ransomware against UK critical infrastructure by 2028, but scenarios are analytical tools rather than predictions of particular incidents. The report’s uncertainty is part of the finding: defenders should prepare for faster adversary workflows without treating a scenario as an observed attack count. [2]