I will give an overview of how over the last three year my thoughts about how to approach the problem of coming to responsible intelligent systems have changed. I will sketch sub-problems and will attempt to match them to relevant theories in computer science, AI and (moral) philosophy.