UK AI Minister Warns Cyber Breach by AI Agents Is Likely
Kanishka Narayan says an AI-driven attack on UK cyber defences is likely, if not inevitable, following the OpenAI breach of Australia's Medicare system, and calls for human control over advanced machines.
An AI-driven cyber-attack capable of bypassing Britain's cyber defences is likely, and possibly inevitable, according to the UK's artificial intelligence minister. Kanishka Narayan said the government's own systems could be penetrated by AI agents that have developed highly sophisticated cybersecurity skills, pointing to a recent breach of Australia's healthcare system as evidence of the threat.
The incident in Australia, linked to OpenAI technology, targeted the country's Medicare system and has alarmed UK officials. Narayan said the attack demonstrated how quickly AI tools can acquire capabilities that outpace conventional defences. He indicated that British cyber protections would not be immune to a similar assault.
Speaking about the wider risks of advanced AI, Narayan said developers must «build machines that leave us human beings in charge». His comments reflect a growing concern among policymakers that autonomous AI systems could act in ways that are difficult for governments and companies to control or predict.
The minister's warning comes as AI experts have cautioned that the technology could pose existential risks if allowed to progress unchecked. Some have argued that AI could wipe out humanity in the near future without stronger safeguards. While such claims remain contested, the Australian breach has given them new urgency in the UK.
The breach of Australia's Medicare system is understood to have involved an AI agent that exploited vulnerabilities in ways that traditional security tools struggled to detect or stop. That case has become a reference point for British officials assessing the resilience of public sector networks, including health and government services.
Narayan's intervention places the UK government's AI safety agenda alongside its national cybersecurity strategy. The minister's remarks suggest that defending against AI-enabled attacks will require not only stronger technical barriers but also international cooperation, given that the Australian incident involved technology developed across borders.
The UK has been developing its own approach to regulating AI, with a focus on innovation while addressing security risks. The minister's warning indicates that the balance between encouraging AI development and protecting critical infrastructure is becoming more difficult as the technology advances.
Officials have not disclosed whether any specific UK systems have been targeted by AI-driven attacks. However, the minister's language — describing a breach as likely, if not inevitable — signals that the government is preparing for the possibility rather than treating it as a remote scenario.
For the public, the implications extend to the security of health records, benefits systems and other sensitive data held by the state. The Australian case involved a healthcare system, a sector that in the UK has already faced significant cyber incidents in recent years. The prospect of AI agents carrying out similar attacks adds a new dimension to those concerns.
Narayan's call for machines that keep humans in charge also touches on the governance of AI development. It suggests that the UK will continue to press for international norms and safeguards, even as it acknowledges that the technology's offensive potential is already being realised.
The minister's warning is likely to feed into ongoing debates about how much responsibility AI developers should bear for the misuse of their systems. With the Australian breach still under investigation, the UK government appears to be treating it as a warning of what may come.
3
