Britain's critical national infrastructure should expect a ransomware attack that uses AI-powered malware within the next two years, the UK's National Cyber Security Centre (NCSC) warned on Tuesday. The judgement anchors the agency's new State of the Cyber Nation report, and it comes with a blunt message for boards: the era of "tech on the cheap" is over.
The NCSC, part of the signals intelligence agency GCHQ, published the warning alongside a formal threat assessment, Artificial intelligence and cyber intrusion, which projects the AI-enabled threat to the UK out to 2028. It judges it "highly likely" that AI-powered malware will have been used in a ransomware attack on UK critical national infrastructure (CNI) by then, using the formal probability language that UK intelligence assessments are bound to.
Old weaknesses, found at machine speed
The agency's core argument is not that AI invents new attacks. In remarks prepared for the report's launch, NCSC chief executive Dr Richard Horne says the country faces "a real prospect of an unprecedented rise in the number of cyber incidents" because vulnerabilities "that have been tolerated for far too long can now be discovered and exploited at a scale that was once unimaginable."
The assessment's key judgements spell out what that looks like by 2028:
- It is highly likely that Russia, China and other nation states will routinely use AI-enabled cyber capability to pursue geopolitical goals, raising the scale and speed of sophisticated attacks on the UK.
- It is almost certain that many systems will remain vulnerable to AI tools that find and exploit weaknesses, letting attackers take less care to protect an access because replacing it costs them little.
- It is likely that attackers will turn AI capability on known vulnerabilities in legacy CNI technology, and on the new attack surface created when CNI operators deploy AI tools of their own.
The report sketches three scenarios rather than one forecast. In the baseline case, AI-driven discovery has sharply increased the number of zero-day vulnerabilities found and defensive tooling has shortened patch windows, but defensive capability in UK CNI still lags behind offensive capability. In the pessimistic case, frontier AI proliferates to everyone from state actors to novices, attribution gets harder, and the NCSC rates it a "realistic possibility" that an automated actor no longer under its operator's control could self-replicate and attack at will. The optimistic case depends on CNI operators replacing ageing systems and on a workforce trained in AI and security. The NCSC stresses that these are plausible trajectories, not predictions, and that "the prospect for surprise is high."
On specific states, the assessment judges that Russian actors have highly likely adapted and retrained open-source AI models rather than building their own, while China has made significant strides despite lagging on chip quality.
Horne's fix: invest, disconnect, go beyond the baseline
Horne's prescription is mostly unglamorous. He urges organisations to fix underinvestment in their technology, stop leaving systems connected to the internet when they do not need to be, and treat Cyber Essentials as the starting point rather than the finish line, building further resilience in line with their size, complexity and risk. Leaders who do not, he warns, will have to explain to investors, customers and suppliers why "food isn't on shelves, lights aren't on, businesses have gone under."
The government is pairing the warning with money. The Chancellor last week announced £60 million for the first phase of a national Cyber Shield, £70 million for a National Exposure and Real-Time Vulnerability (NERV) Centre to find weaknesses across critical infrastructure and the economy, and an AI Cyber Security Centre of Excellence in Manchester run by the NCSC. The Cyber Security and Resilience Bill and a forthcoming National Cyber Action Plan sit alongside them.
The report extends a run of NCSC warnings on the same theme, from its statement on frontier AI models acting on their own to its alert on exposed industrial control systems. More on the UK's threat picture is on our United Kingdom country profile. The common thread is that the exposed, unpatched system is the real risk, and AI mainly shortens the time it takes somebody to find it.
This briefing is provided by IntelFusions for informational and defensive purposes only. It is based on sources assessed to be reliable at the time of writing, and analytic judgments carry the confidence levels indicated. Indicators of compromise are defanged; re-arm them only in controlled environments. IntelFusions is not affiliated with the organizations named and makes no warranty as to completeness or accuracy.