I gave an overview on recent advances in causal representation learning and possible relations to causal incentives analysis in AGI safety.