The UK’s AI minister, Kanishka Narayan, set out a package of measures against risks from increasingly powerful AI on Monday 12 October, in a speech at Imperial College London’s science and tech incubator, Base One, the government said [1]. The measures include putting frontier AI to work on cyber defence and considering stronger protections through the Cyber Security and Resilience Bill regime [1].
CONFIRMED (government text): from the GOV.UK press release published at 11:14 BST on 12 October 2026 by the Cabinet Office, the AI Security Institute, the National Cyber Security Centre and DCMS. Proposals on the Bill are under consideration, not law.
What he said
“I want Britain to become one of the hardest countries in the world to attack using artificial intelligence,” Narayan said, according to the release [1].
“We will use every power at our disposal. We will legislate wherever necessary. And we won’t look to law alone,” he said [1].
He also addressed AI developers directly: “And to the companies developing the world’s most powerful AI systems, my message is: Listen when your scientists warn of serious dangers. Investigate unexpected behaviour. And if you cannot adequately control the risks, do not release the system.” [1]
The release names no company and gives no funding figure.
The measures
Frontier AI for cyber defence. The government said it will put “frontier AI to work strengthening cyber defences, building on pilots that have already identified and addressed over 1,000 vulnerabilities” [1]. It said the two pilots “helped identify and fix over a thousand potential cyber vulnerabilities across government systems” [1].
Possible changes to the Cyber Security and Resilience Bill regime. The release says “Stronger protections against emerging AI threats” are “to be considered through the Cyber Security and Resilience Bill regime, such as how serious AI incidents and near misses can be identified and reported” [1]. The government “will explore strengthening parts of the Cyber Security and Resilience Bill and the wider regime”, which “includes AI loss of control risk, where autonomous AI systems evade human oversight” [1]. It “will look at how near-miss incidents can be identified and reported” [1]. These are proposals: “The detail of any measures will be worked through with parliamentarians, as well as technical experts and the operators who would need to deliver them.” [1]
The Financial Times reported on the same day that “The UK government is drawing up plans for compulsory safety reporting and monitoring of frontier AI” [3]. “Compulsory” is the FT’s word; the government’s release does not use it.
Incidents involving AI agents. The release says work is underway with the National Cyber Security Centre and the AI Security Institute “to look at how incidents involving AI agents – which can perform complex tasks with little human supervision – should be responded to” [1].
AISI evaluations. Narayan “confirmed that scientists at Britain’s AI Security Institute are now conducting rigorous evaluations of the latest generation of advanced AI models, following work with leading AI developers and the US Government,” the release says [1]. It does not name the developers or models.
AI verification. “More details on support for UK AI verification efforts will be set out in the coming months.” [1]
Cyber minister
Cyber Minister Liz Lloyd said organisations face “an undeniable reality – it’s now business-critical to ensure you’re on the front-foot when it comes to cyber security.” [1]
Before the speech
The Guardian previewed the speech overnight [2]. It also quoted Geoffrey Irving, “a former chief scientist at the UK’s AI Security Institute”, who “told the Guardian”: “The AI companies do not have security that would prevent an adversarial country from stealing the models.” The Guardian added: “OpenAI says it has multiple protections against theft of its models, including encryption and continuous monitoring.” [2] TSN has not checked either account. Several points in the Guardian’s preview do not appear in the government’s release, and TSN has not used them.
What is not known
- The full speech text: the government has published a press release, not a transcript.
- Which AI developers and models AISI is evaluating.
- What, if any, changes will be made to the Cyber Security and Resilience Bill, and when.
Sources
- GOV.UK, “Britain steps up national defences against growing AI risks”, Cabinet Office, AI Security Institute, National Cyber Security Centre and Department for Culture, Media and Sport, published 12 October 2026, 11:14 BST. https://www.gov.uk/government/news/britain-steps-up-national-defences-against-growing-ai-risks
- Robert Booth and Dan Milmo, “UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts”, the Guardian, published 00:01 BST, 12 October 2026 (preview). https://www.theguardian.com/technology/2026/oct/12/uk-strengthen-cyber-defences-minister-ai-firms-safety
- Financial Times, 12 October 2026. https://www.ft.com/content/ba28b7cd-293e-4eee-a073-ad81aaefced3
Related stories
- White House Reportedly Reverses Course, Restoring UK AI Security Institute’s Early Access to US AI Models
- White House Super Intelligence Force Says AI Companies Must Report Security Incidents and Remedy Harm; Axios Says Enforcement Is Not Spelled Out
- EU Tech Chief Virkkunen Warns of Under-Tested AI Models and Says Brussels Has Sent Information Requests to OpenAI, Anthropic and Chinese Firms
- Microsoft CEO Satya Nadella Says AI Models Should Be Treated Like Insider Risks
