When utility meets ethics: A stakeholder perspective on agentic information systems delegation

Abstract

Organizations increasingly use agentic information systems (agentic IS) to achieve performance-driven objectives. However, agentic IS raise ethical concerns, such as fairness, bias, autonomy, transparency, responsibility, and privacy. Over 70 % of the stakeholders, such as users, have expressed concerns about the various business applications of agentic IS. However, limited research has examined how stakeholders evaluate utility versus ethics in delegating tasks to agentic IS. Our results show that stakeholders assess agentic IS for delegation using a combination of rules-based and outcome-based ethical thinking, which values both AI utility and ethical considerations. Moreover, when agentic IS embeds ethical capabilities, stakeholders view both agentic IS and delegating tasks to them as ethically sound. Furthermore, stakeholders consistently prefer agentic IS with human oversight over fully autonomous systems. These insights help organizations optimize efficiency through agentic IS while promoting ethical adoption by respecting stakeholder preferences. This research contributes to the IS literature by integrating ethics into the agentic delegation framework, thereby establishing a foundation for future research in the developing field of agentic IS.