Investigating ethical issues for persuasive systems

Abstract

The ability of discerning good from wrong will be fundamental for autonomous artificial agents, especially when they communicate with humans. In fact, as such agents are becoming more complex and common in our ev- eryday life, the need for an ethical design cannot be avoided, especially when persuasive systems are concerned. Starting from the cognitive tradition based on moral dilemmas, a set of principled guidelines for design and implemen- tation of ethical persuasive systems is proposed, along with some challenges, in particular concerning meta-planning models of ethical reasoning. Further- more, to assess how our guidelines stand in relation to existing persuasion prototypes, we have designed a questionnaire and administered it to their developers. The answers are reported and discussed.