Only a Coward Hides Behind AI? Preferences in Surrogate, Moral Decision-Making

Abstract

The increasing availability of data and computational power in the digital era has transformed many decision-making challenges. Artificial Intelligence (AI) has entered the decision-making stage by taking over more and more tasks previously afforded to humans. The delegation of decision-making authority to AI assigns the technology with the agency to cause moral consequences, which raises concerns about responsibility and accountability. This study explores managers’ willingness-to-delegate moral decisions in a surrogate decision context to AI and the involved employee’s acceptance of such moral decisions made by AI. In a mixed methods approach, we combine quantitative, quasi-experimental results with qualitative insights. Our findings reveal opposing perceptions of the same situation, depending on the individual’s perspective: Whereas the willingness- to-delegate a layoff decision in a surrogate decision context to AI is very low (decision-maker perspective), people affected by the decision have no preference regarding humans or AI in making the decision (decision-affected perspective). Further, we identify the underlying rationales of these distinct preferences along three themes: (1) AI’s capability for this decision based on the technologies’ characteristics and competencies, (2) the particular characteristics of the decision context, and (3) the responsibility for the moral consequences this decision creates.