Representatives of prominent AI firm Anthropic have held meetings in No10 this morning. Along with other AI specialists…
Anthropic’s latest AI model Mythos can identify critical vulnerabilities in a vast array of security systems (including national ones). It can do so with software that is more than 30 years old. Anthropic is keeping the model private for now while it arranges a consortium of firms and partners to shore up security systems before any wider release – this is called Project Glasswing. So far the Russians and Chinese are still behind – No10 should get into those meetings ASAP…
The British government is also attempting to court Anthropic amid its clash with the Pentagon in the US, exploring options for expansion and even a dual-listing in London. Separate AI firm and key Anthropic rival OpenAI has paused a major UK infrastructure project today citing energy costs and regulation. Labour at risk of shooting itself in the foot here…
Danny Kruger has written to the government urging them to get ahead of the Mythos risks by engaging with Anthropic. To fend off “catastrophic cybersecurity risks to the UK”…
Starmer said to Robert Peston this afternoon:
“I’m fed up with the fact that families across the country see their bills go up and down on energy businesses bills go up and down on energy because of the actions of Putin or Trump across the world.”