We are excited to reveal the new phase of The Institute for Ethical AI Alignment & Safety.
Join us in our mission to ensure that frontier AI is safe, aligned and open for people and society.
Since our founding in 2017 we have built a reach of over 100,000 practitioners and a track record of contributions across public and private institutions. This new phase brings a revamp across all of our critical initiatives, and the mission that started it continues unchanged.
You can watch the announcement in the reveal video, and read on for what the new phase covers.
The mission
The Institute is an independent research institute working to ensure that frontier AI is safe, aligned and open for people and society. That work has always sat at the intersection of three things that are usually kept apart: the regulation that sets the obligations, the infrastructure that makes those obligations implementable, and the practitioner network that has to live with both.
The new phase is a consolidation of those three tracks rather than a change of direction.
Regulation and policy
On the regulation front we now count over 30 policy contributions across the European Commission, the United Nations and beyond, with several of our recommendations adopted across EU and UK policy.
That record is documented in full on the policy page, alongside the frameworks that translate the obligations into something a team can actually run against.
Infrastructure and open source
On the infrastructure side we have built a long range of open source projects and initiatives. Some are now hosted with The Linux Foundation, several are referenced by the OWASP Foundation, and across the ecosystem they have gathered over 200 contributors and more than 20,000 GitHub stars.
The full set, including the projects that are actively maintained and the curated production lists, lives on the open source page.
The network
As of today we have built a reach and network of over 100,000 practitioners. This includes a newsletter that now reaches almost 70,000 practitioners across 393 published issues, together with a growing following across our social channels. We are grateful for the support the community has given us across nearly a decade.
The Ethical AI Network is where that community comes together, and the newsletter archive carries every issue we have published.
What has been updated
Several of our key initiatives have been revised as part of this phase:
- The Nine Principles for AI Alignment & Safety, rewritten around the failure domain each principle addresses, the commitment that answers it, and the controls that implement it.
- The State of Production ML Survey, updated with the 2025 results and set side by side with 2024 so the movement is visible.
- The frameworks, including the maturity models and the procurement templates, which we will continue to update through this phase.
Get involved
If this is the work you want to be close to, there are three ways in:
- Join the network to take part in the initiatives directly.
- Subscribe to the newsletter for the weekly research and industry roundup.
- Contact us if you are working on something where we could be useful.
Thank you to everyone who has supported this work since 2017. The next phase starts now.