From Decision-Maker to Decision Architect: Generative AI and the Changing Role of Leadership

Abstract

Generative artificial intelligence (GenAI) is becoming an integral part of managerial work, changing how leaders analyse information, communicate, coordinate, and decide. Much of the existing leadership literature emphasizes productivity effects or debates substitution versus augmentation. We argue that the more fundamental leadership challenge is the reconfiguration of decision authority and the design of decision processes across human–AI arrangements. As GenAI systems become more capable, context-aware, and agentic, leadership becomes less about individual judgement and more about structuring how judgement is distributed across human and AI actors. This conceptual paper asks how GenAI changes leadership decision-making in contemporary organizations. Drawing on research on algorithmic management, the automation–augmentation paradox, AI-mediated leadership, and principal–agent perspectives on AI agents, we develop a framework that positions leaders as decision architects, i.e., actors who deliberately design and govern decision processes rather than merely participate in them. The main contribution of the paper is a typology of four roles GenAI can have in leadership decision making, distinguished by the degree of AI autonomy and the locus of decision authority: AI as assistant, AI as advisor, AI as oversight partner, and AI as delegated actor. For each role, we discuss implications for decision rights, employee autonomy, and managerial expertise. We further argue that the agentic AI highlights classical delegation problems of goal alignment, information asymmetries, and control, turning governance decisions into a core leadership concern. By focusing on design of human–AI decision architectures, the paper contributes to leadership and organization literature by clarifying how GenAI can change concepts of authority, control, and legitimacy in contemporary organizations. We outline a research agenda on AI-augmented leadership capabilities, responsible delegation to AI agents, and the conditions under which GenAI can improve leadership legitimacy.