Seminar Algorithm Credulity: Human and Algorithmic Advice in Prediction Experiments
28 April 2026
Eric Guerci - Professor at Université Côte d’Azur - Nice (France)
- 12:00 PM - 01:00 PM
- Online on Microsoft Teams and in person : Dipartimento di Scienze Aziendali - Via Capo di Lucca n°34 - Sala Seminari 1, Primo Piano, Bologna
- Science & Technology In English
How to partecipate
Free admission subject to availability
Program
The seminar is reserved for the Department of Management’s community. Other interested colleagues can contact Elisa Villani (e.villani@unibo.it).
This seminar examines algorithm credulity, namely the tendency to rely on faulty algorithmic advice without sufficient critical evaluation. Drawing on prediction experiments comparing human and algorithmic advisors, the study shows that participants are more likely to follow the same deficient advice when it is provided by an algorithm rather than by a human. This tendency leads to a 13% reduction in expected earnings. To explain this result, the paper introduces the Algo-Intelligibility-Credulity Model, according to which people are more likely to perceive unpredictable and deficient advice as intelligible when it comes from an algorithm. The findings highlight an important vulnerability in human interaction with automated systems and suggest that misleading algorithmic advice may be especially influential.
Short biography
Eric Guerci is Associate Professor and Maître de Conférences (HDR) in Economics at Université Côte d’Azur and a member of GREDEG-CNRS. He is also Deputy Director of the Laboratoire d’Économie Expérimentale de Nice (LEEN). His research lies at the intersection of behavioral and experimental economics, computational economics, and the study of emotions, with recent work focusing on human responses to algorithmic advice and decision-making in experimental settings.