The Effect of Algorithm Performance Feedback on Negotiator Ruthlessness

Abstract: Algorithms are increasingly integrated into decision-making processes across various domains, yet human’s distrust of information generated from algorithms—termed algorithm aversion—can undermine their potential benefits. We investigate the impact of algorithm-generated performance feedback on negotiation behavior, focusing on how negotiators' perceptions of the source of the feedback (e.g., human-generated vs. algorithm-generated) influence their decision-making and interaction patterns. Two studies showed that negotiators receiving feedback from algorithms exhibit more ruthless behavior, compared to those receiving feedback from human experts, by behaving in a more competitively (i.e., lower Social Value Orientation [SVO]) and making more distributive decisions (i.e., aspirational first offers). Our findings highlight how the source of the feedback meant to improve skills at the negotiation table can inadvertently alter the interaction dynamic resulting in more competitive behavior. Understanding these dynamics is crucial for optimizing the integration of algorithms in decision-making processes and improving their efficacy in collaborative settings.

Keywords: AI-human Interaction, Ruthlessness, Competitive Behaviors, Distributive Negotiations