Features of Moral Consideration for Artificial Entities: A Conjoint Experiment

摘要

The moral consideration of intelligent artificial entities (e.g., robots, virtual personal assistants) is a topic of growing academic interest. Studies have identified a range of features that are associated with the moral consideration of such entities, often in the context of two accounts: mind perception and humanness (i.e., anthropomorphism and dehumanization). The present study brings together and builds on this growing body of research by evaluating the relative importance of the various features of moral consideration. We conducted an online conjoint experiment in which 1,163 participants evaluated 30,238 profiles of artificial entities that randomly varied on 11 features. All 11 features affected the extent to which participants consider it morally wrong to harm an artificial entity. The two most important features were an entity’s capacity for emotion expression and moral judgment. These were followed by emotion recognition, cooperation, and the entity’s body (in particular, having a human-like physical body). Overall, the results provide support for a humanness account of moral consideration: the more human-like artificial entities are perceived to be in their mental, physical, and behavioral characteristics, the more moral consideration they are given. Within the humanness account, the study supports the view that capacities that are associated with the “human nature” dimension more strongly affect moral consideration than capacities associated with the “uniquely human” dimension, but both dimensions positively affect moral consideration.