Ethical implications of verbal disinhibition with conversational agents

This paper presents a reflection on the ethical implications of conversational agents. The reflection is motivated by recent empirical findings showing that, when interacting in natural language with artificial partners, users tend to indulge in disinhibited behaviour, such as flaming, bullying and sexual harassment. The paper then addresses the question whether conversational agents open any ethical issues and whether this new communication context requires the definition of new moral values and principles or could be addressed by ordinary moral norms.