0

UK AI Security Institute (UK AISI)

British government institute (formerly AI Safety Institute) conducting pre-deployment dangerous-capability evals on frontier models.

Visit
Type
non profit
HQ
London, UK
Founded
2023 (announced at the Bletchley Park AI Safety Summit; renamed to AI Security Institute Feb 2025)
Funding
Government-funded (UK Department for Science, Innovation and Technology)
Website
aisi.gov.uk
GitHub
github.com/UKGovernmentBEIS
Twitter
twitter.com/AISafetyInst

Cite

Notes

Only stored in your browser.

Evals
1
Tools
0
Models
0
Papers
0
Boards
0
People
3

Evals

1

AgentHarm: Harmfulness Potential in AI Agents

Assesses whether AI agents might engage in harmful activities by testing their responses to malicious prompts in areas like cybercrime, harassment, and fraud, aiming to ensure safe behavior.

ActiveSafeguards