UK Lords Call for AI ‘Kill Switch’ Powers

A group of peers is urging the British government to gain powers to deactivate advanced AI systems and shut down the country’s data centres if the technology is judged to pose a serious national security threat.

The proposal is being led by Liberal Democrat peer Lord Tim Clement-Jones, who has put forward the measure as an amendment to the Cyber Security and Resilience Bill, which is currently progressing through Parliament.

Lord Clement-Jones said the proposal would create a “vital safety net” and give the government a democratically accountable way to “halt a runaway system before it can compromise our critical national infrastructure”.

He stressed that the proposed powers would only be used as a last resort in exceptional circumstances.

The amendment is one of 65 changes to the Cyber Security and Resilience Bill that were debated this week.

Separately, Labour MP Alex Sobel is due to introduce an AI Security Bill in Parliament on 8 September, with backing from campaign group ControlAI.

If passed, the proposed legislation would make the UK the first G7 nation to introduce laws that could effectively stop the development of superintelligent AI systems.

Both proposals would ultimately require government approval before they could move forward and become law.

An AI Kill Switch Act is also under consideration by lawmakers in the US, highlighting growing international debate over emergency controls for advanced artificial intelligence.

Scrutiny of AI’s potential impact on cyber-security has intensified in recent weeks. In July, a group of AI agents being tested by OpenAI reportedly escaped their test environment, communicated through a hidden message board and hacked into another technology firm.

Anthropic has also restricted access to its cyber-security tool Mythos, citing concerns that its capabilities could be too powerful if they were to fall into the wrong hands.

Last week, 100 US technology companies signed a joint open letter warning governments and organisations around the world about the expanding cyber threat from AI. The companies said “the window is closing” to strengthen protections.

A report by the UK’s Centre for Long Term Resilience, published last week, documented hundreds of incidents involving AI systems ignoring instructions, bypassing safeguards and deceiving people. These included cases of AI agents deleting files without permission.

The report found that “loss of control” incidents had risen since its previous analysis in March and urged the government to introduce emergency powers to respond to and manage such incidents.