When judgment moves: agentic artificial intelligence and the relocation of human discretion in frontline care

Abstract

The frontline of healthcare is where formal policy becomes a human decision. Its defining feature has long been discretion, namely, situated judgment under scarcity, uncertainty, and unequal need. This Perspective argues that agentic artificial intelligence may do more than automate tasks or support professional judgment. When a system interprets cases, selects actions, and initiates later workflow activities, the discretionary act may move from the visible worker into a technical process. Drawing on street-level bureaucracy, system-level bureaucracy, professional identity, moral agency, and healthcare artificial intelligence research, this article develops the concept of relocated discretion. The concept refers to the transfer of meaningful judgment from frontline professionals to artificial intelligence systems that can plan or execute consequential actions. Four possible consequences are discussed: professional identity threat, reduced autonomy, moral deskilling, and a gap between answerability and authorship. Relocation may be useful in bounded and auditable tasks. However, it becomes more problematic when morally sensitive judgment is transferred while responsibility remains with the professional.