UK AI Security Institute
The UK AI Security Institute (AISI), formerly the AI Safety Institute, is a research organisation within the UK government's Department for Science, Innovation and Technology that evaluates frontier AI systems for risks to national security and public safety. It conducts pre-release model evaluations and red-teaming with major AI labs such as Anthropic, Google, and OpenAI, and develops open-source evaluation tools to advance AI safety research globally.
“Errors in automated alignment research are likely to be harder to identify than the human baseline.”
“Britain's AI Security Institute — 'staffed by former OpenAI and Google researchers, weapons experts, epidemiologists, and code breakers' — has already found that leading models 'could be coaxed into giving instructions for bioweapons and cyberattacks.'”