US Reverses Ban: UK AI Safety Institute to Test Models from Anthropic, Google, OpenAI

The White House has reversed its September decision to block the United Kingdom’s AI Safety Institute from accessing new artificial intelligence models, according to The Decoder. The agency will now resume early access to frontier systems developed by Anthropic, Google DeepMind, and OpenAI, restoring a channel for pre-release safety evaluations that had been suspended.

The Daily Telegraph reports that the reinstated access applies specifically to models from the three named technology companies. The White House had previously told these firms to stop sharing their models with the British agency. This directive led Anthropic to restrict access to its new Claude Mythos 5.1 model. The reversal allows the UK AI Safety Institute to once again get early access to these models for safety risk testing.

Diplomatic Pressure and Policy Reversal

The reversal came after diplomatic pressure from UK AI Minister Kanishka Narayan. In September, the Trump administration had instructed the companies to cease sharing their models with the AI Safety Institute. Narayan’s efforts led to the block being lifted. Narayan plans to announce new AI safety measures in an upcoming speech, including stricter reporting requirements for AI-related security incidents involving critical infrastructure.

International Context and Security Warnings

Prime Minister Andy Burnham had previously pushed for international AI regulation at the UN General Assembly. At the same event, President Trump warned that "globalist" organizations should stay out of American AI development. Additionally, heads of MI6, MI5, and GCHQ had warned ministers that AI technology was empowering hostile actors and terrorists. These warnings highlight the security concerns surrounding AI development and the need for robust safety measures.

Defender's checklist

Small-business owners should monitor official government announcements regarding AI safety standards and regulatory changes. It is important to stay informed about how international cooperation on AI safety might impact the tools and services your business relies on. Understanding the landscape of AI regulation helps in anticipating future compliance requirements or shifts in available technology.

What to do and how to stay safe: Anthropic

• Monitor official channels for updates on AI safety regulations and international cooperation agreements.
• Review your current use of AI tools to ensure they align with emerging safety standards.
• Stay informed about potential changes in access to frontier AI models and their implications.
• Educate your team on the importance of AI safety and the role of international testing agencies.

Step-by-step guide: Implementing AI Usage Policies at Work: A Step-by-Step Framework

General security guidance from the InfoSec Ledger newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Which AI companies are affected by the White House's decision?

Anthropic, Google DeepMind, and OpenAI are the three companies whose models will again be accessible to the UK AI Safety Institute.

Why was the UK AI Safety Institute blocked from accessing models?

The Trump administration had instructed the companies to stop sharing models with the agency in September, citing concerns about globalist organizations.

What is the role of the UK AI Safety Institute?

The agency tests AI systems for safety risks before they ship, ensuring they meet certain safety standards.