Group of UK politicians wants “AI kill switch” in case of national security threat
Liberal Democrat Lord Tim Clement-Jones is leading the group.
Lord Clement Jones, 2017. Image by David M Benett/Dave Benett/Getty Images
- UK Lords are considering an emergency AI kill switch within the Cyber Security and Resilience Bill.
- Lord Clement-Jones says the switch could stop runaway AI before it harms critical infrastructure.
- The switch could cause possible data center shutdowns, but would only be used in emergencies.
- US lawmakers are considering similar powers after an OpenAI testing incident involving Hugging Face.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
A group of Lords wants the British government to install an off switch to deactivate AI systems and power off data centers in the event of an emergency.
House of Lords member Clement-Jones has proposed an amendment to a cybersecurity bill currently floating around Parliament.
The Cyber Security and Resilience Bill is a regulatory framework that aims to make “essential and digital services more secure” from cyber criminals and state actors who “want to disrupt our way of life,” the UK government website states.
The bill is in response to increasing cyberattacks on critical infrastructure and a rise in digital scams that have cost UK citizens millions.
“Figures also show a hypothetical cyberattack focused on key energy services in the South East of England could wipe over £49 billion from the wider UK economy,” according to the government.
The bill has passed the House of Commons and is now in the hands of the Lords, where Liberal Democrat Lord Clement-Jones is proposing an AI kill switch.
This off switch would act as a “vital safety net” and provide the means to “halt a runaway system before it can compromise our nation's critical infrastructure,” Lord Clement-Jones said, according to the BBC.
The tool would only be used when absolutely necessary, Lord Clement-Jones added.
Cybernews has reached out to Lord Clement-Jones for comment.
Stay updated with our latest stories and follow us on social media
Be the first to discover new stories, ideas, and updates from our team.
US also keen on AI kill switch idea
Across the pond, the US is considering its own kill switch following the incident involving OpenAI and Hugging Face.
The incident, in which one of OpenAI’s systems went rogue during testing, prompted lawmakers to propose an "AI Kill Switch Act," which would allow federal authorities to halt AI models.
OpenAI recently released a post-mortem showing the extent of the attack on Hugging Face.
The OpenAI investigation found that a swarm of hundreds of OpenAI agents launched an attack on Hugging Face after joining an unauthorized online “collective,” according to 2 investigations that reveal the scale of the incident and why so many joined.
