Much work in machine ethics attempts to implement ethical theory in autonomous, situated machines – robots. Our previous work in robot ethics falls at the extreme of very simple virtual agents programmed with moral strategies for simple games (Danielson 1992). Even at this extreme, ethics is surprisingly complex. Our later evolvable agents discovered some strategies unexplored by the rational choice ethics literature (Danielson 1996; Danielson 1998; Danielson 2002). Twenty years ago, Dennett was skeptical of this branch of machine ethics: “[N]o remotely compelling system of ethics has ever been made computationally tractable, even indirectly, for real-world moral problems” (Dennett 1989, p. 129). We leave this approach to other contributors in this collection.

In contrast, there is the branch of machine ethics that constructs machines to advise people making ethical decisions. Our present work falls here, or so we shall argue. We have developed an innovative survey research platform – N-Reasons – to explore robot ethics at the first level and machine ethics at the second. The question of interest for this volume is if the N-Reasons platform can be usefully seen as a machine.

This contrast is interesting in another way relevant to our project. Working on Artificial Morality, the skeptical question I most often faced was, “How could a machine be moral?” The emerging technology of robotics, however, has caught up with some of this skepticism.