Who's who · Governments and regulators
AI Security Institute (UK)
United Kingdom: government body that tests frontier AI models for security and safety risks
The UK government's AI testing institute, the first of its kind, founded as the AI Safety Institute in 2023.
Why it matters: It has pre-release testing access to models from leading US labs.
Key facts
- Launched on 2 November 2023, around the AI Safety Summit at Bletchley Park, as the AI Safety Institute.Source: en.wikipedia.org, as of Nov 2026
- Renamed the AI Security Institute in February 2025, announced by Peter Kyle at the Munich Security Conference.Source: en.wikipedia.org, as of Nov 2026
- As of February 2026 had pre-release testing agreements with Anthropic, Google and OpenAI, and maintains the open-source Inspect evaluation platform.Source: en.wikipedia.org, as of Feb 2026
- Was moved under the new Office for the Prime Minister and Cabinet in a July 2026 government reorganisation.Source: en.wikipedia.org, as of Jul 2026
Our latest stories
No stories about AI Security Institute (UK) in our recent coverage yet. New ones appear here as they publish.