Redistributing Decision Authority in AI-Augmented Organizations: Toward a Theory of Managerial Accountability

Abstract

The redistribution of decision authority across human and technological actors raises fundamental questions about the substantive executability of managerial accountability in AI-augmented organizations. Although organizational authority, human–AI delegation, and control–accountability alignment theories explain the distribution of decision influence and responsibility, they provide a less differentiated account of how managers mobilize the capacities required to fulfill formal accountability obligations. Addressing this explanatory gap, the study develops Managerial Accountability Executability Theory (MAET) through the conceptual integration of organizational authority, accountability, human–AI delegation, and algorithmic control perspectives. The proposed theory introduces the accountability execution gap, defined as the discrepancy between formal accountability demands and effectively accessible interpretive, intervention, and remedial capacities. Five theoretical propositions specify the roles of capacity complementarity, organizational accessibility, decision consequence severity, and procedural approval in shaping substantive managerial answerability. By distinguishing integrated, accessible distributed, and fragmented executability, MAET explains why distributed decision control does not necessarily undermine accountability. The framework advances a capacity-based understanding of managerial accountability and provides theoretically grounded directions for empirical investigation and organizational governance.