Point de vue Quand l’IA active les biais et la déresponsabilisation des décideurs : risque pour les individus et les organisations

Abstract

This article examines how the integration of artificial intelligence (AI) reshapes managerial decision-making by activating two often underestimated dynamics: the loss of account ability among actors and the exacerbation of cognitive biases. Based on a literature review, it shows that the use of AI, presented as a mere decision-support tool, transforms human agency to the point of reducing individuals to passive validators. At the organizational level, the adoption of AI induces processes of algorithmization that reconfigure routines, governance, and power relations, with technical dependencies likely to obscure decision criteria and erode discretionary judgment. The article proposes a set of recommendations addressing both AI design and management. Taken together, these insights shed light on the conditions for AI-assisted decision-making that preserve human agency and responsibility, while emphasizing the need for a close coupling between design, practices, and organizational mechanisms.