Software engineers rarely think about the safety of their products, but as the world becomes more connected, involves more machine learning and deploys more sensors the potential for harm only increases. How should an engineer think about these challenges? What approaches can be added to the development lifecycle in order to factor in ethics into the process. How has software gone wrong in the past and how have the organizations in charge iterated to prevent misuse of their products? [Content Warning: This talk discusses some aspects of software in the defense space]