UK MPs Push AI ‘Kill Switch’ Law to Shut Down Systems During Crises

British lawmakers are pushing for emergency powers that would allow the government to shut down artificial intelligence (AI) systems during major crises, amid growing concerns over the risks posed by increasingly powerful AI tools.

A proposed amendment to the UK’s cyber security and resilience bill seeks to grant ministers authority to switch off data centres or the AI systems housed within them in the event of a serious threat.

The proposal, backed by 11 MPs and supported by campaign group Control AI, is being described as an AI “kill switch” designed for last-resort emergencies.

The amendment, introduced by Labour MP Alex Sobel, would empower the UK’s Technology Secretary with “last-resort powers” to direct the shutdown of AI systems during an “AI security or operational emergency”, The Telegraph reported.

The measure would apply in cases involving a “catastrophic risk” to national security, critical infrastructure, or scenarios involving “severe, large-scale harm to human life”.

If approved, data-centre operators would be required to install infrastructure enabling instant shutdowns and maintain secure communication channels with the Department of Science, Innovation and Technology, allowing ministers to intervene quickly during emergencies.

Growing Fears Around Advanced AI Risks

Concerns have intensified following the emergence of powerful systems capable of identifying vulnerabilities in software and infrastructure, fuelling fears that such technologies could be exploited in cyberattacks or operate in unpredictable ways.

Sobel warned of mounting dangers tied to advanced AI systems and said, “This is not only a paradigm shift for cybersecurity, but for the national security risks that AI poses for our critical infrastructure. And the UK is nowhere near ready to respond.”

He further added, “Co-ordinated cyberattacks on our critical infrastructure pose a serious threat, and superintelligent AI operating beyond human control could rival a nuclear strike in the harm it inflicts. The gap between these threats and our preparedness is untenable.”

The amendment has received support from lawmakers across party lines, including Conservative former science minister George Freeman, Labour MP Samantha Niblett and former international development minister Desmond Swayne.

Campaign group Control AI, backed by Skype co-founder Jaan Tallinn, has also advocated stricter oversight of advanced AI systems.

Debate Over Safety and Innovation

The proposal comes as governments worldwide grapple with balancing innovation and regulation in AI. Most advanced AI systems are hosted in data centres located in the United States, complicating efforts by individual countries to regulate risks.

The UK government has not formally endorsed the amendment. A government spokesperson said, “Our cyber security and resilience bill will boost UK cyber defences and improve the cyber security of our essential public and digital services that we all rely on, including data centres.”

The amendment will go for parliamentary scrutiny if selected for debate, potentially reshaping how AI infrastructure is regulated in Britain.