The United Kingdom has warned that artificial intelligence could significantly increase the scale and speed of cyberattacks against essential infrastructure, with an AI-powered ransomware attack on critical national infrastructure considered highly likely by 2028.
The warning was issued by the UK’s National Cyber Security Centre (NCSC), part of the Government Communications Headquarters (GCHQ), in its newly published 2026 annual review and assessment of artificial intelligence and cyber threats.
The assessment highlights growing risks to systems supporting electricity, water, transport, healthcare, communications, manufacturing and other essential services.
According to the NCSC, artificial intelligence is not necessarily creating entirely new methods of cyberattacks. Instead, it is making existing techniques faster, more efficient and easier to deploy at scale. AI tools can help attackers identify weaknesses in software, examine exposed systems, automate parts of their operations and develop more convincing social-engineering campaigns.
The agency warned that organisations relying on outdated technology, weak security practices and underfunded digital infrastructure could become particularly vulnerable.
NCSC Chief Executive Richard Horne said organisations had relied for too long on inadequate technology and insufficient investment in cybersecurity. He warned that AI could expose longstanding weaknesses at a speed and scale that businesses and public institutions may struggle to manage.
The report also assesses that Russia, China and other state actors are likely to use AI-enabled cyber capabilities more routinely by 2028 to pursue geopolitical objectives.
China is described in the assessment as having the strongest capability and intent to target UK critical national infrastructure and strategic assets. The report also highlights Russian efforts to penetrate UK military networks and malicious cyber activities attributed to Iran.
These are intelligence assessments by the British cybersecurity agency, not proof that every future cyberattack will originate from those countries. The governments concerned have previously rejected similar allegations of malicious cyber activity.
The NCSC also warned about the growing use of AI by criminal groups. Automated tools could make ransomware operations, vulnerability discovery and targeted phishing more efficient, potentially increasing the number of organisations exposed to attacks.
However, the report makes clear that AI is not only a security risk. The same technology can help defenders identify weaknesses, analyse threats and respond to incidents more quickly. Organisations with mature cybersecurity systems and responsible AI governance may gain stronger defensive capabilities.
The agency has called for greater investment in cybersecurity, better management of software vulnerabilities, stronger supply-chain protection and improved monitoring of digital systems.
The warning comes amid a series of cyber incidents affecting businesses and essential services in different countries. Recent attacks on European retail and supermarket systems have demonstrated how weaknesses in digital networks can disrupt physical operations.
The NCSC’s assessment is therefore a warning about the direction of the cyber threat landscape rather than confirmation of a specific planned attack. Its central message is that organisations must strengthen their defences before attackers gain an advantage from increasingly capable AI tools.