No doubt our AI descendants will commit horrors. It would be nice if we could influence AGI moral development, but it seems rather implausible to me that we'll have any greater power to "align" AGI's than Homo Erectus had over the alignment of Homo Sapiens.