Blaming artificial agents for difficult moral choices in medical contexts -- Study 2

Le résumé fourni par la source

In this follow-up experimental study, we explore how people perceive and judge difficult moral decisions made in a medical context, specifically contrasting decisions made by a human advisor with those made by an artificial intelligent (AI) advisor, and how the advisor type interacts with the types of reason the advisor is motivated by or explicitly states and, critically, whether the stated and motivating reasons align or diverge. Participants will be presented with one critical medical scarcity scenario (providing a ventilator) where an advisor must make a life-or-death decision. In a between-subjects 2x3x3 factorial design, we manipulate advisor type (decisions made by a human advisor versus an AI advisor) and the type of reason (fairness-based, utilitarian, institutional concern-based), both for motivating and stated reasons. Like in Study 1, Participants will report how blameworthy, trustworthy, and empathic they think each advisor is as well as how morally justified the decision was; whether the advisor could have decided otherwise; how much regret does the advisor feels; how much blame the institution deserves.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.