AI, the unexpected attractor for tax and customs administrations



  • Ignacio Miguel Gonzalez UPM
  • Salvador Duart Crespo


Artificial Intelligence, Machine Learning, Simplexity, Fairness, AlgorAI, Automation, Compliance, Equitable Society, Comfort Spheres, Innovation


This paper explores how Artificial Intelligence (AI) is becoming the most influential disruptive technology in Tax and Customs Administrations (TCAs) and how it will inexorably lead to a new paradigm. TCAs will evolve from complex, algorithmically driven, efficiency-based organizations of the twentieth century to new ones characterized by ‘Simplexity’, 'AlgorAI' governance, and automation of rule-based human decisions aimed at compliance. It: i) examines the transition towards AI integration across TCAs; ii) presents a roadmap for AI adoption, emphasizing the balance between technology and human factors; iii) discusses the development of digital "Tax Information Spaces"; and iv) suggests using AI to enhance, not replace, human expertise through "Comfort Spheres." This paper addresses the right balance between Human-In-The-Loop (HITL) and Human-On-The-Loop (HOTL) approaches and recommends progressing towards a new state of Human-Above-The-Loop (HATL).

The paper concludes that the successful integration of AI in TCAs requires a balanced approach that considers both technological advancements and human factors. The study underscores the importance of addressing taxpayer and civil servant reactions to AI, recommending strategies to navigate these challenges and achieve a harmonious integration of AI in tax and customs processes.

Author Biography

Salvador Duart Crespo

Mr. Duart is a seasoned Customs and Trade Facilitation specialist with over 25 years of experience, spanning roles in private, public, multilateral, and international organizations





How to Cite

Gonzalez, I. M., & Salvador Duart Crespo. (2024). AI, the unexpected attractor for tax and customs administrations: TAMING THE LOOP. Review of International and European Economic Law, 3(5). Retrieved from