Would people react positively or negatively if organizational decisions were made by AI instead of by traditional decision agents (i.e., human managers)? Whereas high-profile stories covered in news media suggest that the use of AI in organizational decision-making was viewed as unfair and received negatively, recent surveys showed that using AI in decision-making was seen as fair and received positively by the general public. The current research reconciles these conflicting observations.

We conducted studies in the context of dispute resolution, culpability judgment, and performance evaluation to examine how individuals react to AI-made (vs. human-made) decisions. The studies showed that when the decision outcomes were favorable to the self (e.g., getting a bonus), whether AI or human was the decision agent did not matter much. However, when the decision outcomes were unfavorable (e.g., getting punished), AI was seen as fairer than human. Importantly, the decision recipients were more likely to accept and internalize the negative outcomes. However, a brief reminder about AI biases in decision-making attenuated the halo effects of AI.

From the perspective of organizations, the use of AI can help legitimize decisions, especially unpopular ones. From the societal perspective, the fairness halo that is blindly attributed to AI by the unsuspecting public might propagate or even exaggerate existing biases. With the increasing use of AI in organizational decision-making, we need to use Artificial Intelligence (AI) intelligently by increasing our awareness about the strengths and pitfalls of AI-made decisions.