Assaf Harpaz (Georgia) presented Taxing AI at the Mizzou Law Tax Policy Colloquium on October 22:

Artificial intelligence (AI) is changing the world and is introducing numerous challenges to legal and regulatory frameworks. These tensions are highlighted in federal income taxation, which broadly serves three principal goals: revenue generation, redistribution, and regulation of taxpayer behavior. To achieve these goals, the U.S. tax system disproportionately relies on the taxation of individual labor income and payroll, rather than capital taxation. 

The widespread integration of AI is poised to reshape the labor market, potentially transforming both the sources of income (from humans to robots) and the character of income (from labor to capital) subject to tax. Moreover, new AI models can create value and operate autonomously, challenging the traditional conceptions of legal personhood. Under current U.S. law, even the most advanced AI models are not directly subject to the income tax regime, as they are neither individuals nor business entities.

This Article promotes a functionalist methodology to analyze the risks that AI poses to federal income taxation. It argues that AI’s autonomous features and its capacity to disrupt the labor market undermine the income tax system’s basic goals. The Article offers paths for reform, focusing on questions of taxpayer entity and personhood. In doing so, the Article advances the scholarly debate beyond capital income and “robot taxes.” It proposes that AI’s autonomy and a looming labor market transformation could drive change in the federal income tax doctrine.