Peter Barnett

Peter was part of the first MATS cohort, in 2022, working with Leo Gao. Peter now works at the Machine Intelligence Research Institute on the Technical Governance Team. He focuses on strategy and governance plans to prevent catastrophic outcomes from advanced AI. He has also previously worked on the limitations and uses of AI evaluations. Prior to working on AI governance, Peter worked on technical alignment at MIRI and on reward learning at Center for Human-Compatible AI at UC Berkeley.