Self-driving vehicles (SDVs) will sometimes face moral dilemmas and have to decide whether to protect their passengers at all cost or minimize the number of casualties. There might be two kinds of crash algorithms programed in SDVs to deal with these dilemmas. A selfish algorithm means that SDVs will protect their passengers at all cost. A utilitarian algorithm means that the moral decision of SDVs is to minimize all social losses. We collected and compared participants’ responses to the selfish and utilitarian SDVs in terms of deontological evaluation (i.e., moral righteousness of the moral decision of these two SDVs and the adoption of these two SDVs), perceived benefit and perceived risk of these SDVs, and behavioral intention to use these SDVs. Compared with the utilitarian SDVs, participants (N = 580) perceived greater benefit from the selfish SDVs and reported a greater intention to use the selfish SDVs. Participants perceived non-different risk from these two vehicles and evaluated the righteousness of the moral decision of both SDVs at the same level (through deontological evaluation). Deontological evaluation, perceived benefit and perceived risk were three significant predictors for behavioral intention and after controlling for them, vehicle type still exerted a direct influence on behavioral intention. Our results imply that developing and deploying the selfish SDVs (vs. the utilitarian SDVs) seems to be more socially acceptable.

CICTP