New development: I chose to let the machine choose—the rise of <i>‘</i> AIcratism <i>’</i>

This article conceptualizes ‘AIcratism’, an emerging phenomenon that occurs when public employees, despite possessing formal decision-making authority, inappropriately defer to AI, thereby transferring agency to algorithms. Whereas existing research has largely emphasized the strategic and policy dimensions of AI adoption in the public sector, this work shifts the focus to the micro-level of daily decision-making, focusing on organizational implications. Drawing on a sociomaterial perspective, it reflects on how discretion is exercised, and potentially undermined, when human actors inappropriately rely on algorithmic suggestions. The conceptualization of AIcratism in this article advances the literature by illuminating how algorithmic systems can subtly reshape discretion and accountability in public organizations. The article offers a framework for analysing quasi-automated bureaucracies and outlines avenues for future research on human–AI collaboration, appropriate reliance, and organizational design in AI-supported public service delivery.