Managers' Trust in AI Expert Recommendations

Abstract

Managers today face increasing pressure to integrate AI-generated insights into their decision-making processes. This study examines how managers’ trust in expert recommendations varies depending on whether the advice is from a human or AI expert. This study also examines the extent to which heightened levels of accountability influence the relationship between managers trust in a human or AI recommendations. Using a 2x2 between-subjects experiment, upper-level accounting students acting as CFOs were randomly assigned to receive a recommendation from a human expert or an AI expert, with half of each group also receiving a heightened level of accountability. Results indicate a significant difference in trust between expert source, with human recommendations being trusted more than AI recommendations, supporting Hypothesis 1. Contrary to Hypothesis 2, accountability manipulations did not significantly influence trust in either recommendation source. These findings highlight persistent algorithm aversion among decision makers and suggest that simply increasing accountability does not mitigate lower trust in AI experts.