The Authority Allocation Problem in AI-Augmented Organizations

Abstract

This paper examines how judgment authority becomes difficult to locate in AI-augmented organizations. As artificial intelligence systems increasingly participate in institutional decision-making-recommending, classifying, prioritizing, approving, rejecting, and acting-the right to decide migrates from formal organizational structures into system configuration: thresholds, routing rules, defaults, and workflows. The prevailing instruments of AI governance address system properties and outputs but leave unanswered who, within the organization, holds legitimate decision authority. Through analysis of implicit authority, the paper identifies three recurring patterns across AI-augmented organizations: (1) Authority Allocation Gaps, (2) Escalation Gaps, and (3) Accountability Continuity Gaps. These constitute the paper's primary findings. The observed dynamics can be interpreted as failures of institutional boundary formation-a reading for which the construct of Governance Decision Boundaries may serve as an analytical lens. The paper considers implications across government, financial services, healthcare, and agentic systems, and suggests that AI governance may be understood, at least in part, as an authority allocation problem. AI Disclosure Generative AI tools, including ChatGPT, Claude, and Codex, were used to support drafting, editing, and language refinement. All conceptual development, analytical interpretation, argumentation, and final editorial decisions were made by the author.

Keywords: Decision Design, Governance Decision Boundaries, AI Governance, Authority Allocation, Judgment Architecture, Human-AI Systems, Organizational Decision Making, Artificial Intelligence, Institutional Governance