Attending Labs is an AI reliability and safety lab. AI is beginning to improve itself, and it is entering the real world faster than it can be verified. We build the reliability that lets the world trust AI, and the safety research that keeps self-improving systems under human control.

We are a team from Stanford, Oxford, Columbia, and Google. We have published research on uncertainty for generative AI, and brought high-performance, on-device AI to billions of users on Chrome.

Backed by

Email address Subscribe