This paper introduces the Task Assignment Paradox (TAP): the structural mismatch between how organizations currently deploy artificial intelligence — assigning tasks to AI within human-centered workflows — and what AI's correct definition requires. Building on the Language-Intelligence-Process framework established in the Transition Trilogy and the definitional foundations of Paper Four, this paper argues that human-centered judgment workflows are not merely suboptimal but constitutively incompatible with AI as defined by the Law of Intelligence Propagation (LIP). Evidence converges across three independent traditions: the historical record of industrial transformation (Thompson, 1966), cognitive science research on partial automation (Bainbridge, 1983), and contemporary enterprise performance data (McKinsey, 2025; NVIDIA, 2026). The conclusion is not that AI should replace human workers, but that workflow structures themselves must be redesigned with AI as the primary judgment subject. The Seoul Metropolitan Government's 2026 transition to AI-Readable administrative documents is presented as the earliest large-scale government-level instantiation of this structural shift.
Kyungae Ahn (Wed,) studied this question.