fraþimind, understanding
pioneering the research of international defence.
mission
Frathi brings together researchers, engineers and policy specialists from allied nations to build artificial intelligence that keeps people safe. We study how AI can help in defence, and how to keep human judgement in command.
research
The problems we're working on first. Each one is pinned on the map on the home page.
Independent testing and red-teaming of AI systems before they are trusted with real decisions.
Machine-speed detection and response, and protection for AI models against poisoning, evasion and theft.
Tools that help commanders and analysts see their options faster, while the decision stays with people.
publications
Our first papers, reports and datasets are being written now. They'll be published here, openly wherever we can.
Notes on AI safety and evaluation
partnerships
We're open to working with universities, industry and governments. To start a conversation, get in touch.
Partnership programme in the works.
Contact uscareers
We're building a team of researchers, engineers and policy specialists. New roles will appear here as they open.
No positions yet.
contact
For research partnerships, press and general enquiries, email us and we'll get back to you.
contact@frathi.com
error 404
The page you're looking for isn't on our map. It may have moved, or the address may be mistyped.
--.----° n --.----° e