Algorithmic Tax Power and Constitutional Safeguards: Global Perspectives on AI, Bias, and Digital Tax Justice

Abstract

The growing use of algorithms and artificial intelligence (AI) systems in tax administration poses significant challenges to constitutional principles that have historically constrained the state’s power to levy taxes. This paper analyses how the automated exercise of taxing authority—here referred to as the “artificial taxing power”—can reproduce biases, obscure decision-making processes, and undermine fundamental rights such as legality, equality, ability to pay, and due process. Based on doctrinal, constitutional, and empirical analysis, the study argues that the Brazilian Constitution, from 1946 to the present, provides a solid foundation for the creation of a Digital Taxpayer Bill of Rights. Comparative experiences (EU, US, China and Latin America) are examined, the risks of generative models (LLMs) are discussed, and a governance agenda is proposed for digital tax justice, including legal, institutional, and technical safeguards to ensure transparency, explainability, and human oversight.