In this talk, Magdalena explains different research directions in technical AI safety. The talk is for you if you already believe that AI safety is an important problem, and want to get a better understanding of current technical research.
https://www.youtube.com/watch?v=fKZahM44KQA