On AI:AM

Geoffrey Irving

Cofounder & Chief Scientist, Sequent

Geoffrey Irving is cofounder and Chief Scientist of Sequent, a new nonprofit alignment research organization that launched on this episode. He previously served as Chief Scientist of the UK's AI Security Institute, led language model teams at DeepMind, and did early RLHF work at OpenAI. He authored the foundational 'AI safety via debate' (2018) and works on scalable oversight, formal verification, and the theoretical foundations of alignment.