Friendly Aritificial Intelligence is simply a theory that AI would be beneficial to humans and provide positive change. While the ethics of machines are exploring the ideas behind how a machine should act, current research on friendly AI is focusing on how to make, or teach, AI applications how to behave in a manner that would fit with the machine ethics.
This research also includes how to keep AI applications to show reserve and essentially to be controlled so they do not start behaving inappropriately or incorrectly. A good example of this would be the Laws of Robotics that Isaac Asimov created in his science fiction book, "I, Robot".