AI Risk and Vulnerability Alliance

AVID is the flagship effort of the non-profit organization AI Risk and Vulnerability Alliance (ARVA).


As ARVA, our mission is to empower communities to recognize, diagnose, and manage vulnerabilities in AI that affects them.


Our vision is to work across disciplines to make AI safer for everyone by exposing vulnerabilities in the AI ecosystem. We build tools for practitioners and communities to interrogate the growing number of models capable of producing material harm to our world. Together we operationalize AI risk management to protect the future of our world.


Our work as ARVA—and AVID—is guided the following values.

Builders First

Our community leads by example. We prioritize making things work over telling others how they should work. We come up with imperfect solutions to collectively iterate and improve upon.


To make the impact we strive for we must hold ourselves to a high standard. While we may work in a scrappy fashion we build things with practical usage and interoperability in mind.


We keep watch for new vulnerabilities and work together to cut through the hype around new AI products and innovations.


Our work demands we take action and speak against the growing reliance and proliferation of AI by corporations who put profit before all.


In all situations we treat each other with dignity. Different cultures, experiences, and circumstances create both conflict and the opportunity to come together. Only by appreciating those who are different from us can we find better ways forward.


We reach out to those who hold different values, and have different needs, to find mutually beneficial solutions. Our effort to bridge gaps and break down barriers leads us to build tools and conduct research inclusive of individuals from all walks of life.

Board of Directors

Subho MajumdarNathan ButtersBrian Pendleton
Rumman ChowdhuryKush VarshneyBorhane Blili-Hamelin