The third speaker for the 2022/2023 PIBBSS Speaker Series. This was recorded in June 2023.
AI alignment research can be classified as either developing different components of an aligned AI system (e.g. reward design and generalization) or enabling more effective work on the components (e.g. through improving interpretability or theoretical understanding). This talk will give an overview of research directions in each of these areas and how alignment research at DeepMind fits into this framework.