AI-based Decision-Making : Not the Decision-Maker but the Outcome’s Favourability Determines the Perception of University Topic Allocations

Kurzfassung/Abstract

Over the past few years, artificial intelligence (AI) has increasingly taken over decision-making processes in our daily lives. Prior studies observed differences in the perception and evaluation of AI-based, algorithmic decision-making versus human decision-making that diverged between decision-making contexts. At present, especially educational contexts remain underexplored. In this study, we investigated the influence of the decision-maker (AI vs. human) on fairness perception, trust, and emotional responses in participants (N = 329) who are allocated university course topics. Importantly, all topics were allocated according to the same principles after students indicated their preferences. However, lecturers in the human condition pretended to make the decision themselves based on the information provided, whereas lecturers in the AI condition explicitly stated that the AI decided. Furthermore, we manipulated whether the corresponding decision-making process was explicitly communicated as fair or whether no comment was made on the process’ fairness. We further assessed how favourable students rated the outcome of the allocation process. Against our hypotheses, Bayesian evidence indicated that neither the decision-maker nor whether the decision-making process was communicated as fair had an impact on students' fairness perception, trust, or emotional responses. The central determinant of students’ evaluations of the university course topic allocation process was the favourability of the outcome. Given that an AI can better optimize allocations according to students’ preferences than human decision-makers, we argue that these findings advocate for a broader implementation of AI-based decision-making in the context of allocation decisions at university.