The Buck Stops There: The Impact of Perceived Accountability and Control on the Intention to Delegate to Software Agents

Abstract

Software agents with the ability to recommend actions, aid decision-making, and actually make decisions are becoming increasingly common. In many situations, users now choose whether or not to delegate tasks to these agents. While some research has examined software agents, relatively little is known about the factors that influence the intention to delegate decisions to them. An experiment was used to examine the influence of perceived accountability, extent of control, and trust in the agent on the intention to delegate a travel arrangement decision. Users were more likely to delegate to agents that gave them greater control by requiring them to approve the agent’s recommendation before the decision was completed than to agents that performed the task autonomously without intervention after it was delegated. Contrary to expectations, intention to delegate increased as perceived accountability increased. Participants may perceive delegation as a means to shift blame from themselves to the agent and thus mitigate risk resulting from potential negative decision outcomes.