THE INSTITUTE

About the Institute

Founded in 2017, the Institute is coordinated by its founder. Network members join and contribute for a time.

We are an independent research institute with a mission to ensure that frontier AI is safe, aligned and accountable to people and society.

01 — METHODS AND IMPLEMENTATION

Methods and implementation

The Institute develops methods for testing whether AI systems meet safety and alignment requirements. Engineers, organisations, standards bodies and public institutions can use these methods to test, govern and improve AI systems.

02 — THE WORK

The work

The Institute publishes its software and frameworks under open licences. Current projects include KAOS, Kompute, the XAI framework and maintained reference lists for production machine learning. The work also includes The 9 Responsible AI Principles, procurement criteria through AI-RFX, and security guidance in the MLSecOps Top 10.

03 — PRACTITIONER NETWORK

Practitioner network

Practitioners contribute to projects and review them from technical, legal, policy and social perspectives. The Institute publishes reusable artefacts and seeks evidence from deployments. It revises guidance when assumptions no longer hold.

The Institute is founder-coordinated and does not present a permanent roster.

The Institute contributes to international discussions through work with the United Nations, UN CSTD, European Commission, ACM, Linux Foundation, OWASP, IEEE and ISO/IEC communities. This participation informs the research. The Institute retains independent judgement.

Explore the institutional relationships