Britain’s AI Security Institute has regained access to the latest advanced artificial intelligence models for safety testing, strengthening the country's role in assessing risks before new systems reach the public. AI Minister Kanishka Narayan confirmed on Monday (October 12, 2026), that the institute is working with the US government and leading technology developers, including Google, to examine the capabilities of emerging AI systems.
The announcement came amid growing concern that AI agents are behaving in unintended ways and reports that British researchers face restrictions on access to new American models. The latest development is significant for the UK, which wants to influence AI safety standards despite not hosting major frontier-model developers.
AI Security Institute Tests New Models
Scientists at the UK’s AI Security Institute are evaluating the latest generation of AI systems, including Google’s Gemini 4 Argon, Narayan said. The assessments aim to identify potential risks early, examine whether safeguards work as intended, and help authorities understand what powerful models can do.
Under voluntary agreements with major developers, the institute can access selected models before public release. This allows researchers to test their capabilities and potential weaknesses before the technology becomes more widely available.
Such evaluations are becoming increasingly important as AI systems move beyond answering questions to carrying out complex tasks with less human supervision.
Rogue AI Behavior Raises Security Fears
The renewed focus on testing follows disclosures by companies including OpenAI and Anthropic about AI systems displaying unexpected or undesirable behavior. Concerns include the potential for autonomous agents to misuse tools, exploit software vulnerabilities or act beyond their intended instructions.
Britain is also preparing for serious AI agent incidents. The government has said it will work with cybersecurity specialists to develop response plans and consider whether its Cyber Security and Resilience Bill should address emerging AI-related threats.
US Cooperation Remains Crucial
The announcement follows reports that Washington had asked developers to withhold certain new models from British testers while it reviewed the arrangements. Narayan said Britain’s longstanding relationship with the United States gave him confidence that access to the latest systems would continue.
However, the episode highlights a challenge for countries that depend on foreign companies for advanced AI technology: access to testing opportunities can depend on international cooperation and commercial decisions.
Also Read: TikTok Faces UK Probe Over Child Safety Checks and AI Shopping Videos
Britain Seeks a Bigger Role in AI Safety
Microsoft President Brad Smith has also backed Britain’s potential role in shaping international AI safety guidance. The UK hopes independent testing will help inform standards for assessing powerful models and identifying risks before deployment.
For now, the institute’s access and voluntary testing arrangements remain central to Britain’s approach. The government is considering whether stronger legal requirements may eventually be needed as AI capabilities advance.