Abstract

This chapter argues that machine ethics has a deontic bias narrowly focusing on the concerns of social morality. This bias distorts the machine morality debate by promoting an impoverished view of moral theory, resulting in three issues. First, it weakens any claims arguing for the possibility of machine morality – the idea that machines can be moral subjects, not just instrumental objects. Second, it overlooks potentially rewarding lines of inquiry for future research. Third, as an interdisciplinary field, it does moral philosophy a disservice by portraying moral theory in an oversimplified light. The example of supererogation is used to demonstrate this point. While supererogatory acts are both familiar and undeniably moral in nature, the idea of machine supererogation (that machines could go beyond the call of duty) is conceptually problematic. This exemplifies the problem with assuming human moral concepts can accommodate artificial intelligence without substantial revision.