From Computation to Irreversibility: Why AI Cannot Replace Judgment

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Author: Xu, Lucas Xiaochun
Format: Recurso digital
Published: Zenodo 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866901149369499648
author Xu, Lucas Xiaochun
author_facet Xu, Lucas Xiaochun
contents <div class="el-p"> <h2>Abstract</h2> <p>This paper does not introduce a new idea. It reveals the <strong>boundary condition</strong> under which all prior judgment theories become necessary.</p> <p>Recent advances in artificial intelligence have led to widespread claims that human expertise can be systematically replicated, distilled, and eventually replaced. This paper challenges that assumption by introducing a fundamental distinction between <em>computation</em> and <em>irreversible judgment</em>. While AI systems excel at compressing, reproducing, and scaling past patterns (low-entropy operations), they fundamentally fail in domains characterized by <strong>irreversibility, path dependence, </strong>and <strong>unquantifiable consequence</strong>.</p> <p>We propose a <strong>three-layer "Judgment Stack" model</strong> that explains how AI reorganizes enterprises: (1) computable execution, (2) verifiable judgment, and (3) irreversible judgment. We argue that AI-driven transformation is not merely a productivity upgrade but a thermodynamic restructuring of organizations toward lower entropy states. In this process, human roles are not eliminated uniformly but stratified: replaceable, compressible, or amplified.</p> <p>The paper concludes that AI cannot replace judgment at the highest level—not due to technical limitations alone, but due to the ontological nature of decision-making under irreversibility.</p> </div>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19786741
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle From Computation to Irreversibility: Why AI Cannot Replace Judgment
Xu, Lucas Xiaochun
AI Transformation
irreversible judgment
decision theory
entropy reduction
organizational design
human-AI interaction
risk governance
cognitive architecture
<div class="el-p"> <h2>Abstract</h2> <p>This paper does not introduce a new idea. It reveals the <strong>boundary condition</strong> under which all prior judgment theories become necessary.</p> <p>Recent advances in artificial intelligence have led to widespread claims that human expertise can be systematically replicated, distilled, and eventually replaced. This paper challenges that assumption by introducing a fundamental distinction between <em>computation</em> and <em>irreversible judgment</em>. While AI systems excel at compressing, reproducing, and scaling past patterns (low-entropy operations), they fundamentally fail in domains characterized by <strong>irreversibility, path dependence, </strong>and <strong>unquantifiable consequence</strong>.</p> <p>We propose a <strong>three-layer "Judgment Stack" model</strong> that explains how AI reorganizes enterprises: (1) computable execution, (2) verifiable judgment, and (3) irreversible judgment. We argue that AI-driven transformation is not merely a productivity upgrade but a thermodynamic restructuring of organizations toward lower entropy states. In this process, human roles are not eliminated uniformly but stratified: replaceable, compressible, or amplified.</p> <p>The paper concludes that AI cannot replace judgment at the highest level—not due to technical limitations alone, but due to the ontological nature of decision-making under irreversibility.</p> </div>
title From Computation to Irreversibility: Why AI Cannot Replace Judgment
topic AI Transformation
irreversible judgment
decision theory
entropy reduction
organizational design
human-AI interaction
risk governance
cognitive architecture
url https://doi.org/10.5281/zenodo.19786741