Tax Theory and Feral AI

Abstract

Federal income tax law and theories that undergird it have not fully addressed the status of non-human earners, like animals or artificial intelligences. Scholarship on artificial intelligence and taxation primarily has focused on the taxation of AI’s owners, on whether AI itself should be taxed despite being owned by someone else, or on the philosophical question of taxing sentient AI. Thinking about tax in the context of AI that is neither owned by a person nor sentient, referred to here as “feral AI,” allows us to identify human biases in theoretical justifications of the income tax and to consider whether they are features or bugs. Specifically, this essay uses non-sentient feral AI as a vehicle to explore the benefit principle and welfarist theories, and suggests ways in which those ideas may produce less than ideal results in the context of taxpayers whom heuristics may cause us to dehumanize. The goal of this essay is not to suggest that feral AI should be subject to an income tax. Rather, it asks whether re-examining the tax policy canon in the context of feral AI reveals anything about human-centricity in tax law. The essay concludes that heavy reliance on human preferences in both the benefit principle and welfarist theories produces a poor fit when human preferences are weak, such as disposition of the last-earned dollars of the ultra-wealthy, or when human preferences are delegitimized by the government, such as those of marginalized groups like undocumented workers. Adopting a non-human point of view in tax policy, then, can inform our thinking about taxation outside the boundaries of what humans perceive to be the norm.