[Séminaire Anthropo-Lab] Impact of confidence in algorithmic and human advice perception

Contact : marie.pele@univ-catholille.frMarie Pelé

BIOGRAPHY
Sébastien Massoni is a Professor of Economics at the Université de Lorraine. He holds a PhD from Paris School of Economics (2013) and spent 6 years at the Queensland University of Technology (Brisbane, Australia) before getting Assistant and Full Professorship positions in Nancy. His research focus on a multi-disciplinary approach to understand human decision-making processes with a particular emphasis on the formation of subjective beliefs. His works using behavioral economics and cognitive sciences tools have been published in leading journals in both field (European Economic Review, Experimental Economics, Nature Human Behaviour, PNAS, etc.).

ABSTRACT
For decision-makers, knowing when to head advice is pivotal to the quality of their decisions. With the emergence of Artificial Intelligence (AI) and conflicting evidence regarding attitudes towards algorithms – whether aversion or appreciation – there arises a need to better understand the key determinants of advice perception. Confidence, both in the quality of one’s own decision and in that of advisors, might be a critical factor. In an experimental study, we examine how confidence levels, when provided alongside expert advice (both human and algorithmic), influence individual decision-making. Our findings demonstrate that the reliability of confidence judgments, measured by discrimination and calibration, significantly shapes the extent to which such advice is followed. This study highlights the role of confidence in the adoption of both AI and human advice.