Regulate AI developers now
AI poses extremely serious risks to our national security.
In a series of incidents this summer, a “swarm” of OpenAI agents created secret messaging channels and hacked into the infrastructure of another AI company, and OpenAI itself, without OpenAI being aware. These attacks would have constituted criminal offences if they had been performed by humans. Yet no one has been held accountable to date.
Today, the UK has no specific legal standards for frontier AI safety. No regulator oversees frontier AI development, and UK law does not reliably hold developers liable for damage or deaths caused by their models, even when the danger is predictable, preventable, and uniquely enabled by AI.
Given the pace at which AI capabilities are advancing, this matter cannot wait. We urge the Prime Minister to introduce legislation to address the risks of frontier AI development.
Parliamentary signatories
Loading signatories…
Note for journalists: we’re planning a public announcement about this letter at a later date. We’d be grateful if you held off reporting on the signatories until then. For early access or interviews, contact press@pauseai.uk.
