Governance

AI Security Institute (AISI)

UK expert body for AI security, capability testing and red‑teaming.

What it triggers

  • Under HCWS462, the UK AI Security Institute evaluates national security risks from frontier AI developers.source
  • The UK AI Security Institute conducts capability testing, red-teaming, risk characterisation and research focused on national-security and criminal-misuse risks.source

Definition

The renamed successor to the AI Safety Institute, established within the Department for Science, Innovation and Technology to conduct capability testing, red‑teaming, risk characterisation and research focused on national‑security and criminal‑misuse risks; it is an expert advisory and operational body rather than a primary statutory regulator.

Not to be confused with

AI Safety Institute

The AI Safety Institute is the former name before it was renamed and rebranded as the AI Security Institute.

Australian Artificial Intelligence Safety Institute (AISI)

The Australian Artificial Intelligence Safety Institute is an independent institute in Australia, unlike the UK AI Security Institute which operates within the Department for Science, Innovation and Technology.

Questions and answers

What is the AI Security Institute?

The AI Security Institute is the renamed successor to the AI Safety Institute, established within the Department for Science, Innovation and Technology. It is an expert advisory and operational body that conducts capability testing, red-teaming, risk characterisation and research focused on national-security and criminal-misuse risks.

What does HCWS462 direct the AI Security Institute to do?

Statement HCWS462 directs the UK AI Security Institute to evaluate national security risks from frontier AI developers. The policy took effect on 2025-02-14 and is currently in force.

Is the AI Security Institute a regulator?

No, it is an expert advisory and operational body rather than a primary statutory regulator.