MIT researchers have developed a control framework that enables robots to understand what it means to help or hinder one another and incorporate social reasoning into the tasks they are accomplishing.