Artificial Intelligence
UK AI safety institute regains model access
The White House reversed an earlier block on the UK AI Security Institute, which will again get early access to models from Anthropic, Google DeepMind and OpenAI after diplomatic pressure.
- The White House reversed its decision to block the UK AI Security Institute from accessing new AI models.
- The institute checks AI systems for safety risks before release and was to regain early access to models from Anthropic, Google DeepMind and OpenAI, according to The Daily Telegraph.
- In September, the Trump administration had told the companies to stop sharing their models with the institute, and Anthropic then limited access to Claude Mythos 5.1.
- AI Minister Kanishka Narayan was said to have applied diplomatic pressure, and he planned to announce new safety measures in a speech.
The White House reversed its earlier decision to block the UK AI Security Institute from accessing new AI models. The institute checks AI systems for safety risks before release, and it was to regain early access to models from Anthropic, Google DeepMind and OpenAI, according to The Daily Telegraph.
In September, the Trump administration had told the companies to stop sharing their models with the institute. Anthropic then limited access to its Claude Mythos 5.1 model.
The block was lifted after diplomatic pressure from AI Minister Kanishka Narayan, the article said. Narayan was due to announce new AI safety measures in a speech, including stricter reporting requirements for AI-related security incidents involving critical infrastructure.
The article also said Andy Burnham had earlier called for international AI regulation at the UN General Assembly. It added that Trump warned there that globalist organisations should stay out of American AI development, and that the heads of MI6, MI5 and GCHQ had warned ministers that AI was helping hostile actors and terrorists.
Named in this story
People
- Kanishka Narayan
- applied diplomatic pressure and planned new safety measures
- Andy Burnham
- had pushed for international AI regulation at the UN General Assembly
- Trump
- warned against 'globalist' organisations influencing American AI development
Companies
- Anthropic
- had its models shared with the institute restricted in September
- Google DeepMind
- was among the model providers affected by the restored access
- OpenAI
- was among the model providers affected by the restored access
Organisations
- UK's AI Security Institute
- was due to regain early access to new models and tests systems for safety risks
- MI6
- was one of the agencies whose heads warned ministers about AI risks
- MI5
- was one of the agencies whose heads warned ministers about AI risks
- GCHQ
- was one of the agencies whose heads warned ministers about AI risks
Governments and agencies
- White House
- reversed its earlier block on the institute's access
- Trump administration
- told the companies to stop sharing their models with the institute
Products and systems
- Claude Mythos 5.1
- had access restricted by Anthropic after the September order
How the source tells it
The article was a straightforward news report with little overt emotive framing.
No emotive framing was found in the original.